Book

ATULISMS

Abstract

ATULISMS is the governed memorial of every word Atul Butte said on camera. Four invariants hold across every chapter:

Body

PREFACE
I. THE SIGNATURE LINES
II. ON DATA
III. ON OPEN SCIENCE
IV. ON DRUG REPURPOSING
V. ON GARAGE BIOTECH
VI. ON MEDICINE
VII. ON AI AND SCALABLE PRIVILEGE
VIII. THE PERSONAL STORIES
IX. ON MENTORSHIP AND CAREER
X. THE MAN THEY REMEMBERED
XI. THE NUMBERS
XII. THE ARC
XIII. THE CLAN
TRANSITIONS — Connective Prose

PREFACE

inherits: hadleylab-canonic/CONTENT/BOOKS/ATULISMS/CHAPTERS references: [I-25, G-1]

Atul Butte — computational biologist, data scientist, and physician — died on June 13, 2025, and this book compiles forty-eight YouTube transcripts and 239,612 words of his recorded talks into governed prose. Commissioned by his wife, Tarangini Deshpande, and edited under CANONIC governance so every quote traces to its source. The conventional read is that a memorial is a biography. The actual rule is that the man's own words speak for themselves.

Abstract

Atul Butte was a computational biologist, data scientist, and physician who changed how genomics interfaces with clinical medicine. He died on June 13, 2025. This book is not a biography. It is a compilation — every talk, every lecture, every keynote he gave on camera, distilled into prose. Commissioned by his wife, Tarangini Deshpande, and edited under CANONIC governance so that every quote traces to its source. His words speak for themselves.

Foreword

"Friends who've known me for decades were stunned when I told them that just 5 seconds after meeting him, I knew I was going to marry this man."

"He proposed to me by taking one of my manuscripts, editing it in PostScript, and changing a part to say 'Will you marry me?' He came over and said, 'Look, they got this wrong in your paper.' I got all ready to be outraged."

"After his diagnosis, we started a new ritual. We would clink our mugs and say, 'Cheers, another morning.' Even from the hospital, he insisted on continuing this ritual, determined to savor every moment we had together."

"Atul's name carries deep meaning. It means incomparable, one of a kind, unique. And in every aspect of his life, Atul truly lived up to that name."

"So, Atul, wherever you are today, cheers. Here's to another morning."

This book is not a biography. It is a distillation. Every word Atul ever said on camera — compiled, indexed, and published. I called Dexter in December 2025 and asked him to preserve the words. Not a memoir. Not an interpretation. The man's own voice, in his own cadence, organized so that anyone who heard him speak even once can hear him again.

Gini knew the man. Marina learned the science. Together they frame everything in between.

— Tarangini (Gini) Deshpande, PhD | Atul's wife | NuMedii CEO | Commissioner of this book | Celebration of Life, June 21, 2025

Preface

I met Atul when I was interviewing for the BMI PhD program. He was giving a job talk at the retreat in Asilomar. I was blown away by his charisma, by his excitement. Before that, I was doing research in comparative genomics — which is very cool — but I was truly amazed that we could use the same skills, informatics, to help patients with disease. That was the whole field of translational bioinformatics that he was pitching at the time.

At the beginning of the school year — September 2006 — I wrote him an email asking for a rotation in his lab. I liked it so much I did two quarters of rotations. About a year later, I asked to join his group.

Nearly everything I know about doing science I learned from Atul. He taught me how to ask questions, how to think big, how to mentor and bring collaborative teams together. He taught me the importance of communication and telling a story. He really elevated people around him, making them feel good about themselves and the work that they do, which in turn inspired them to grow. He was incredibly loyal and dedicated to his work and his family, and his energy was contagious.

After my PhD I spent a few years at Pfizer, and he was instrumental in bringing me back to academia — first to Stanford for a year, and then I started my lab at UCSF in 2015. We continued the drug repurposing work we started together as student and mentor. From gene expression reversal on microarrays to single-cell Alzheimer's drug discovery published in Cell. Nineteen years. From one email asking for a rotation to running the institute he built.

He was an incredible person, a visionary scientist, a mentor. He was a devoted husband, a father, and a very dear friend to many of us. His career was defined by his unwavering belief in the power of data to drive discovery, equity, and better health outcomes for all.

His loss is deeply felt, but his influence endures in the lives he touched, the data he unlocked, and the futures he helped shape.

I'd like to dedicate this talk — and this book — to Dr. Atul Butte.

— Marina Sirota, PhD | Professor, UCSF | Director, Bakar Computational Health Sciences Institute | Adapted from the PSB TERI Lecture, January 2026

Compiler's Note

Gini and Marina are the two women who knew him best. Gini knew the man — five seconds, a PostScript proposal, mugs clinked every morning through the ICU. Marina learned the science — nineteen years, from one email asking for a rotation to directing the institute he built. Their voices open this book because they earned it. They frame everything. Where this matters is the compilation method: because every Atul Butte quote in this book traces to a sourced transcript under CANONIC governance, the memorial is an auditable record of his own words rather than an interpretation of them.

I am honored to have known him for the short time I did. He had an exponential effect on me and my career. The innovation speaks for itself. And this book — literally — speaks for itself. These are his words.

I first met Atul as a resident at Stanford University Hospital when I attended a job talk he was hosting to recruit Dennis Wall. Although the room was full of scholars, it might as well have been only Atul, Dennis, and myself in that room for me that day. While his reputation preceded him, I was struck that he knew who I was — or at least pretended to know. Soon thereafter, I joined his lab, and it changed the trajectory of my life. The day he met me, he asked if I was crazy to be in residency with three children at Stanford. I soon dropped out and joined his lab. He taught me to believe in my unique talents and follow my passion. He showed me my worth.

I will never likely find another MD/PhD who taught himself to code that looks any more like me. We shared the rarest profile in medicine — physician-scientists who write software. Not methods for methods' sake, but code that ships, code that changes practice, code that moves discoveries to patients. He saw that in me before I saw it in myself. He dwarfed my dreams and changed my perspective for the grander after thirty minutes. What an inspiration to all like me.

Writing grants with Atul was exhilarating. Winning them was even more.

In October 2013, he assigned me to lead the R01 for the Butte Lab with Carlos Bustamante and Mike Snyder — interpreting variation in human non-coding genomic regions. "$500k direct per year for 3 years. We have to go in for this." He reviewed my sections while babysitting two kids at the Exploratorium: "I read through the whole thing again and edited. I think it's great!" In 2014, he pointed JDRF at our work: "Thanks guys... this might end up being something like free money for us... JDRF really wants to work with us." Aim 1: "using Dexter's approach." In June 2015, he emailed the whole lab: "Congratulations to Dexter! Your first grant of MANY... :-)"

Then the iMessages started.

Feb 16, 2016. Atul: "Dude! What did the program officer say???? Did you really get the grant??" Me: "Yes! It's great that it's BD2K. Badge of crowdsourcing Big Data honor."

Feb 24, 2016. Me: "Check ur email!" "Marcus grant." Atul: "Wow wow wow! Holy SHIT!! Congrats!!!"

Apr 5, 2016. Me: "Does everyone get one of these?" Atul: "Hell no! I've never seen that before! That's awesome!"

Jun 23, 2017. Me: "BOOM!" (CrADLe U01 funded) Atul: "Super happy for you! Well done!"

For many of us who had the privilege of being mentored by Atul, he carved out time for monthly 1:1 meetings which were amongst the most exhilarating professional experiences ever for me. Not that I've ever done cocaine, but I'd imagine it's like how I felt after meeting with him. If I ever thought my dreams were big, he easily dwarfed them and changed my perspective for the grander after only thirty minutes.

In May 2019, when UCF was recruiting me, Atul told them: "If they don't successfully recruit you now, it is highly unlikely they will EVER be able to recruit someone else like you." That's who he was. He elevated you and then told the world.

The last time we spoke was Christmas 2024.

Oct 15, 2024. Me: "Hi Atul. Let's please reconnect when you have a chance. Just heard from Jane Gibson about AMP. Then I talked to Marina who told me you were sick. I'm so sorry and praying for you and your family."

Nov 28, 2024. Atul: "Happy Thanksgiving Dexter! Wishing you and your family all the very best this season and next year! And yes would love to reconnect soon!"

Dec 25, 2024. Atul: "Merry Christmas!"

Six months later, he was gone.

Atul was not just a pioneer in his field; he defined it. From his early days blending computer science with clinical curiosity, to leading UC Health's innovation efforts, he lived at the bleeding edge of precision medicine. His work in mining public data, developing translational algorithms, and applying machine learning to real-world clinical problems remains foundational. Where others saw noise, Atul saw patterns — insights that saved lives, informed policy, and democratized discovery. He had that rare gift of being both a towering intellect and a deeply human presence. He laughed easily, listened deeply, and cared fully.

He taught me the secret to happiness is to NOT practice. Instead, I built the thing. The governed reuse of real-world data that Atul championed his entire career. And now I present it at the competition named for him. Trained under Atul. Now presenting for Atul. The circle closes.

Though he is no longer with us, Atul's spirit endures in every dataset we question more deeply, every model we train more responsibly, and every mentee we take time to support. He lives on in the infrastructure he built, the leaders he cultivated, and the conversations he started. Let us honor him not just in remembrance, but in emulation. It is both a privilege and honor to be considered part of Atul's Clan. May his memory be a blessing, and his legacy our charge.

This book means he will always be in my lab — the brightest star in the galaxy I am now building. His words. His cadence. His conviction that data is frozen knowledge and you and I can bring the heat. Forty-eight transcripts. 239,612 words. Every phrase he polished over two decades of talks, distilled and indexed so that any of us can sit with Atul one more time.

The chapters that follow are built from forty-eight YouTube transcripts, two memorial recordings, the Atul Butte Mafia WhatsApp group (sixty-six members), LinkedIn tributes, and UCSF obituaries. 958 pages. Every word was said by Atul Butte, about Atul Butte, or in the immediate orbit of his memory. The quotes are sourced. The man's words speak for themselves — they always did.

— Dexter Hadley, MD/PhD | Compiler | San Francisco, February 2026

The Ledger

THE DEXTER-ATUL LEDGER
───────────────────────────────────
607 emails             Jan 25, 2013 — Jul 26, 2020
88 iMessages           Sep 25, 2015 — Dec 25, 2024
66-member Mafia        Jun 14, 2025 — present
1 LinkedIn memorial    381 reactions | 23,527 impressions
695 direct messages    12 years

FIRST:  Jan 25, 2013 — Autism Working Group talk, Stanford
LAST:   Dec 25, 2024 — "Merry Christmas!"

THE GRANT YEARS:
  R01 Bustamante + Snyder    Oct 2013
  JDRF — "free money"        Oct 2014
  K01 Big Data               2014-2016
  BD2K — "Dude!"             Feb 2016
  Marcus — "Holy SHIT!!"     Feb 2016
  CrADLe U01 — "BOOM!"      Jun 2017

THE MOVE:
  "I anticipate a great future for you
   at UCSF" — Mar 2015

THE FAREWELL:
  "Merry Christmas!" — Dec 2024
  Six months later, he was gone.

PREFACE | ATULISMS | CHAPTER 00

Sources

Claim	Source	Link
Atul Butte was a computational biologist and physician who advanced translational bioinformatics	Bakar Computational Health Sciences Institute, UCSF	bakarinstitute.ucsf.edu
Atul Butte died on June 13, 2025, of a rare cancer	UCSF obituary and memorial coverage	ucsf.edu
Marina Sirota directs the Bakar Computational Health Sciences Institute at UCSF	UCSF Bakar Institute leadership	bakarinstitute.ucsf.edu
Single-cell Alzheimer's drug-repurposing work was published in the journal Cell	Cell (Elsevier)	cell.com
The Precision Medicine World Conference hosts the Atul Butte Company Competition	Precision Medicine World Conference	pmwcintl.com

I. THE SIGNATURE LINES

inherits: hadleylab-canonic/CONTENT/BOOKS/ATULISMS/CHAPTERS references: [I-25, G-1]

Across dozens of talks between 2012 and 2025, Atul Butte repeated a handful of signature phrases — "data is frozen knowledge," "a trillion points of data," "it will be a national tragedy" — that became his rhetorical DNA. The conventional read is that these were applause lines. The actual rule is that each phrase was a compressed argument he spent two decades polishing.

The Signature Phrases

These are the phrases Atul repeated across dozens of talks — his rhetorical DNA. Where this matters is repetition: a phrase Atul said in nearly every talk for a decade was not a tic but a thesis he had decided was worth saying again until the field acted on it.

"Data is frozen knowledge"

"Data is power. Data is Revolution. Data is frozen knowledge — and you and I can bring the heat, the light, the energy to melt that data and let the knowledge free."

— TEDMED 2012. The most quoted line he ever spoke.

"A trillion points of data"

"Precisely practicing medicine with a trillion points of data."

— Title of his TEDx, LSI, and UCSF talks. The phrase that became his brand.

"If you want to change the world, you can't just keep writing papers about it"

— Said in virtually every talk from 2015 onward. His justification for starting companies.

"It will be a national tragedy"

"We are spending hundreds of billions of dollars in the United States to collect this data. It will be a national tragedy if we don't use this data to improve the practice of medicine. Period."

— Samsung Catalyst, LSI, NextMed, Danaher, CHIP. His standard opening or closing.

What the Phrases Meant

Where this matters is that each signature phrase Atul repeated was a compressed argument, not a slogan: "everything is connected" was a statistical observation he believed interpersonally, and "one word: hope" summed up an entire research program in a single talk.

"Everything is connected"

"He saw that if you accept any threshold correlation, everything ends up being correlated with everything else. And this led to his phrase, 'everything is connected,' and he believed it very, very deeply, and he meant it interpersonally as well."

— Zach Kohane, Celebration of Life

"One word: hope"

"What do I mean by big data in biomedicine? It's about predicting diseases before they strike, explaining rare diseases that defy experts, finding drugs for diseases that lack attention, making sure we do the right, safe, cost-effective thing for patients. I sum all that up with one word: hope."

— Singularity University, Michigan Omenn Lecture

CANONIC compiles Atul Butte's signature phrases into a sourced, governed record, so the lines he polished across two decades of talks stay anchored to the talks that made them.

ATULISMS | Chapter 1 | The Signature Lines

Sources

Claim	Source	Link
"Data is frozen knowledge" was Atul Butte's most-quoted line, from his TEDMED 2012 talk	TEDMED	tedmed.com
"Precisely practicing medicine with a trillion points of data" was the title of his TEDx talk	TEDx Talks	ted.com
Butte argued that to change the world you cannot just keep writing papers, justifying starting companies	Wikipedia, "Atul Butte"	en.wikipedia.org
"One word: hope" framed big data in biomedicine, delivered at Singularity University	Singularity University	su.org
Zachary Kohane was Butte's mentor at the Children's Hospital Informatics Program	Harvard Medical School, Department of Biomedical Informatics	dbmi.hms.harvard.edu

II. ON DATA

inherits: hadleylab-canonic/CONTENT/BOOKS/ATULISMS/CHAPTERS references: [I-25, G-1]

Across talks from TEDxSanFrancisco to the 2012 Michigan Omenn Lecture, Atul Butte argued that data is not oil but soil, and that 99% of the real work is figuring out the killer question. The data already exists; the United States pays doctors to type the most expensive data in America into systems where, on average, it is never used again. The conventional read is that data is the scarce resource. The actual rule is that the question is the hard part.

Data Is Soil, Not Oil

Where this matters is the metaphor: Atul rejected "data is the new oil" because oil is rivalrous and data is not — a copy costs nothing and grows ideas like soil, which is why he insisted the most expensive data in America must not be left frozen and unused.

Data Is Not Oil — Data Is Soil

"Data is the new oil, they say. I hate that saying. Because oil — if I take that barrel of oil, you can't have it. But data — I can have a copy and you can have a copy and I might do magical things with it and you might do differently magical things with it."

"The saying I like better is: data is the next soil — not oil, soil. You plant ideas and data helps them grow."

— CHIP, Danaher

The Most Expensive Data in America

"I call this data the most expensive data in America now because we pay for doctors to type this stuff in. It'll be a tragedy if we don't use it."

"On average, the average data in these systems is never used again."

— LSI, Danaher

The Question Is the Hard Part

"99% of the work for me and the people who think like me now is figuring out what's the question you want to ask. What's that killer question that everyone's wanted to know the answer to and no one realizes we already have the data to ask and answer that question? That's the hard part. It's not writing the code. It's not cloud computing."

— Singularity, NextMed, Michigan Omenn

Data By Itself Doesn't Do Anything

"We're surrounded by data in medicine but data doesn't do anything by itself. It's stuck. It's frozen in these ponds and puddles."

"Go splashing in those puddles."

— TEDxSanFrancisco

Big Data Comes from Small Packages

"Big data comes from small packages just like this."

— Said in every talk while holding up a gene chip

CANONIC keeps Atul Butte's data philosophy sourced and governed, so the soil-not-oil argument stays anchored to the talks where he planted it.

ATULISMS | Chapter 2 | On Data

Sources

Claim	Source	Link
The "data is the new oil" metaphor is widely cited and contested	The Economist, "The world's most valuable resource is no longer oil, but data"	economist.com
Electronic health record data entry imposes a large documentation burden on physicians	Sinsky et al., "Allocation of Physician Time in Ambulatory Practice," Annals of Internal Medicine 2016	acpjournals.org
A large share of collected clinical data is never reused for analysis	Murdoch & Detsky, "The Inevitable Application of Big Data to Health Care," JAMA 2013	jamanetwork.com
Asking the right research question is widely held to be the hardest part of data science	Provost & Fawcett, "Data Science and its Relationship to Big Data," Big Data journal 2013	liebertpub.com
Gene-expression microarrays ("gene chips") enabled high-throughput genomics	National Human Genome Research Institute, "Microarray Technology"	genome.gov

III. ON OPEN SCIENCE

inherits: hadleylab-canonic/CONTENT/BOOKS/ATULISMS/CHAPTERS references: [I-25, G-1]

In his 2012 TEDMED talk, Atul Butte showed that a high-school student could download more than 30,000 breast-cancer samples as easily as a song on iTunes, calling public big data "retroactive crowdsourcing" by 2,400 labs that do not know they are helping. He gave away his secrets because the unmet need was too large to meet alone. The conventional read is that open science gives away your edge. The actual rule is that the taxpayer already funded the data, so sharing it is the return.

The Open-Science Argument

Where this matters is the taxpayer chain: Atul argued that because public funding already paid for the data, an open repository turns one student's curiosity into more breast-cancer samples than any single lab could ever assemble, and that gift compounds into companies, jobs, and discovery.

Retroactive Crowdsourcing

"When I think of public big data, the term I like to think of is retroactive crowdsourcing. 2,400 labs are there to help you. And those labs don't even know they're helping you."

— Singularity, NextMed

If a High School Kid Can Do It

"A high school kid today that needs to do a science fair project — she can go to this website, type in breast cancer, and now download more than 30,000 digital samples of breast cancer, maybe as easily as she could find a song on iTunes. This kid now has more samples available online than any breast cancer researcher will ever have in his or her lab."

— TEDMED 2012

"If a high school kid can do this, we all can do this. We all should be doing this in every one of our organizations that has anything to do with biomedicine."

— Singularity, Michigan Omenn

The Ridiculous NIH Grant

"Let's pretend I'm writing a brand-new NIH grant. I propose that I can get 3,700 of the best breast cancer labs in the world to share data with me for free. You're already laughing. If I actually wrote that in an NIH grant, you would laugh me out. Yet here it is. Sitting there. Waiting for you."

— Michigan Omenn

I'm Giving Away the Secrets

"I'm giving you my secrets because there's so much unmet need out there. We cannot do them all myself."

— LSI, NextMed

The Priesthood

"A kid who builds a website or writes an iPhone app is doing something that the previous generation needed a priesthood for — a set of elites that knew how to do this."

— TEDMED 2012

The Taxpayer Chain

"The check you write funds the science. The science funds the discovery. The discovery becomes a company. The company gets the jobs. The jobs fund the revolution."

— TEDMED 2012

CANONIC compiles Atul Butte's open-science case into a sourced, governed record, so the retroactive-crowdsourcing argument stays anchored to the talks where he made it.

ATULISMS | Chapter 3 | On Open Science

Sources

Claim	Source	Link
Public repositories let a student download tens of thousands of genomic samples, as Butte described for breast cancer	NCBI Gene Expression Omnibus (GEO)	ncbi.nlm.nih.gov
"Retroactive crowdsourcing" reuses data from labs that did not know they were collaborating	Wikipedia, "Crowdsourcing"	en.wikipedia.org
NIH data-sharing policy requires sharing of taxpayer-funded research data	NIH Policy for Data Management and Sharing	grants.nih.gov
Atul Butte described a high-school student downloading breast-cancer samples in his TEDMED talk	TEDMED	tedmed.com
Software lets individuals build apps that once required a specialist priesthood of elites	Andreessen, "Why Software Is Eating the World," Wall Street Journal 2011	wsj.com

IV. ON DRUG REPURPOSING

inherits: hadleylab-canonic/CONTENT/BOOKS/ATULISMS/CHAPTERS references: [I-25, G-1]

Atul Butte built a method reporters called "match.com for drugs" — reverse a disease's gene-expression signature against a drug's, and you find a new use for an old drug, from imipramine for lung cancer to a public-data-to-acquisition path closed in 24 months. He kept an email from a dying Italian man on his slides as the reason. The conventional read is that a new drug must cost a billion dollars and take 10 years. The actual rule is that the answer is often an approved drug we already have.

Finding Drugs On Purpose

Where this matters is intent: Atul's whole repurposing program was about finding drugs on purpose using public data instead of by accident, so that the imipramine-style match that melts a tumor in a mouse reaches the dying Italian man before the billion-dollar decade runs out.

Match.com for Drugs

"The reporters called this 'match.com for drugs.' What's the famous saying? Opposites attract. I got a disease where this gene goes up and this gene goes down, and I can find a drug that can make this one go down and this one go up — baby, there's a match there! That's with two genes. Imagine 20,000 genes."

— Singularity, NextMed, Michigan Omenn

The Molecules Don't Care

"Not even all the cancers look like each other. Now why did we ever even think the cancers would look like each other — just because the same kind of doctor takes care of them? The molecules don't really care what kind of doctor takes care of the disease."

— TEDMED 2012

Instead of Finding Them by Accident

"Instead of finding these by accident, how about we find them on purpose using public data?"

— Singularity, NextMed, Michigan Omenn

The Italian Man's Email

"Hello, my name is... I'm an Italian man, 48 years old, three children — a girl, a boy, and a 17-year-old, all with the same wife. On June 5th, they told me I have small cell lung cancer. I know I will not live longer than a few months. My family is desperate. Please, I'd like to try this new cure."

"This is why we're in this business. It's an email like this every week or two. We cannot keep saying it's going to cost a billion dollars for each one of these drugs, it's going to take 10 years. That is not what patients want to hear."

— Singularity, NextMed

Neither Side Effect Sounds as Bad as Lung Cancer

"Imipramine — it's an antidepressant... makes you really sleepy and in others it widens your QT interval. Actually, neither of those two side effects sound as bad as having lung cancer — 5% survival rate — and the cancer is melted away, it's gone in this mouse."

— Singularity, Champalimaud

From Public Data to Acquisition

"We went from public data to acquisition in 24 months. Inventors happy. Investors happy."

— NextMed, Michigan Omenn

Turn the Crank

"We turn our crank. We got lots of ideas. Here's a drug. There's a drug. And where we got a lot of interesting ideas isn't the new drug, but the new use for the old drug."

— NextMed (said in every talk)

The Legacy Realized (After)

"I just successfully treated a pediatric patient with an ultrarare genetic disorder by repurposing an FDA approved drug. This specific idea I came up with while in Atul's lab."

— Atul Butte Mafia WhatsApp (120363402783687844@g.us), December 9, 2025. In response to Gini's call to "create a project in Atul's memory." Source: WhatsApp ChatStorage.sqlite, compiled SOP-009.

Inflammatix: Drug Repurposing from Public Data

"Just wish Atul was around to see this."

— Atul Butte Mafia WhatsApp (120363402783687844@g.us), September 30, 2025. On Inflammatix receiving FDA clearance for a sepsis diagnostic — "drug repurposing from public data." Atul's core thesis, realized as a commercial product. Source: WhatsApp ChatStorage.sqlite, compiled SOP-009.

The Cell Paper (After)

"This was truly a team effort."

— Marina Sirota, PhD, Acting Director, Bakar Computational Health Sciences Institute, UCSF. July 22, 2025. Published in Cell — Alzheimer's drug repurposing using the methodology Atul championed. Shared in Atul Butte Mafia WhatsApp. Source: WhatsApp ChatStorage.sqlite + Cell publication, July 2025.

CANONIC compiles Atul Butte's drug-repurposing thesis into a sourced, governed record, so the public-data-to-acquisition story stays anchored to the talks and the patients that proved it.

ATULISMS | Chapter 4 | On Drug Repurposing

Sources

Claim	Source	Link
Computational drug repurposing reverses disease gene-expression signatures against drug signatures	Sirota et al., "Discovery and Preclinical Validation of Drug Indications Using Compendia of Public Gene Expression Data," Science Translational Medicine 2011	science.org
Imipramine and related repurposing candidates were validated in lung-cancer models	Jahchan et al., "A Drug Repositioning Approach Identifies Tricyclic Antidepressants as Inhibitors of Small Cell Lung Cancer," Cancer Discovery 2013	ncbi.nlm.nih.gov
Bringing a new drug to market is often cited as costing around a billion dollars over a decade	DiMasi et al., "Innovation in the pharmaceutical industry: New estimates of R&D costs," Journal of Health Economics 2016	sciencedirect.com
An Alzheimer's drug-repurposing study using this methodology was published in Cell	Cell (Elsevier)	cell.com
Inflammatix received FDA clearance for a host-response sepsis diagnostic	U.S. Food and Drug Administration	fda.gov

V. ON GARAGE BIOTECH

inherits: hadleylab-canonic/CONTENT/BOOKS/ATULISMS/CHAPTERS references: [I-25, G-1]

Atul Butte told audiences that the next Amgen or Genentech would come from a garage, because a million public samples and any mouse model you want are available by credit card, and $10,000 gets your predicted drug tested in mice. He outsourced the rat colonoscopy not to save money but to get the best data in the world. The conventional read is that biotech requires a big institution. The actual rule is that all the tools are already there waiting for you.

All the Tools Are Waiting

Where this matters is access: Atul's garage-biotech argument was that the samples, the mouse models, and the outsourced experiments are all available by credit card now, so the only thing standing between a researcher and the next cancer drug is the willingness to add it to the shopping cart.

The Next Amgen Starts in a Garage

"I fundamentally believe the next Amgen, the next Genentech, is going to come from your garage. Because when you go home from this meeting, you got more than a million of those samples and every mouse model you want available by credit card. What else do you need to launch a biotech company today?"

"I don't mean making meth in your garage — no, no — but I do mean making the next cancer drug, diabetes drug. All the tools are there waiting for you."

— Singularity, Champalimaud, Michigan Omenn

Add to Shopping Cart

"$10,000 gets you your predicted drug tested on these mice — and then once you're ready, amazingly, add to shopping cart."

"I can buy an entire mouse model off the internet with a credit card today. What an amazing time this is."

— TEDMED, Singularity, Michigan Omenn

The Rat Colonoscopy

"Not enough people are laughing, so let me illustrate the geometry involved. This is a colonoscope. This is a rat."

"Nobody on my campus knew how to do a rat colonoscopy. So I'm not just outsourcing to save money. I'm outsourcing to get the best damn data in the world."

— Singularity, Michigan Omenn (his biggest laugh line)

Wisdom of the Crowd

"I don't trust any one of these experiments, but I trust what they show me in common. Wisdom of the crowd."

— Singularity, Michigan Omenn

Companies Are Not the Dark Side

"It is not taboo to talk about companies in academia anymore. It better not be, because this is how we have to get these discoveries to patients. We have no other choice."

— Singularity, CHIP

Slide of Significance

"A lot of physicians use these kinds of slides as a slide of shame. For me, this is a slide of significance."

— Every talk (on his conflict-of-interest disclosure)

CANONIC compiles Atul Butte's garage-biotech case into a sourced, governed record, so the argument that the tools are already waiting stays anchored to the talks where he made it.

ATULISMS | Chapter 5 | On Garage Biotech

Sources

Claim	Source	Link
Amgen and Genentech are founding-era biotechnology companies	Wikipedia, "Genentech"	en.wikipedia.org
Mouse models and reagents can be ordered commercially online	The Jackson Laboratory, mouse model resources	jax.org
Public gene-expression repositories provide millions of samples on demand	NCBI Gene Expression Omnibus (GEO)	ncbi.nlm.nih.gov
Outsourcing specialized experiments to contract labs can improve data quality	Wikipedia, "Contract research organization"	en.wikipedia.org
Conflict-of-interest disclosure is a standard requirement in academic medicine	International Committee of Medical Journal Editors, disclosure of interest	icmje.org

VI. ON MEDICINE

Atul Butte trained at Harvard, did his residency, and saw patients before the data science and the drug repurposing made him famous. Across 48 transcripts spanning 2004 to 2025, the physician kept returning to one frustration: medicine codes a death as "visitation from God," offers 1,500 ways to start a patient on type 2 diabetes meds, and calls everything it cannot explain "not otherwise specified." Medicine appears to be a science of certainty. It is in fact a record of its own imprecision.

The Physician's Complaint

The doctor who looked at ICD codes and saw "visitation from God" as a cause of death is the same doctor who knew that the very phrase precision medicine was an admission. If a 2015 initiative had to name medicine "precise," then everything before it was imprecision. These are the words of the physician speaking — not the data scientist, not the entrepreneur — on the absurdities he watched a billion-dollar health system commit while he saw patients.

Imprecise Medicine

"When we enter an era called precision medicine, it means by definition the last thing we were doing was not precision medicine."

— Champalimaud

"How do we code lung cancer in ICD-9? Left lung or right lung. This is imprecise medicine."

— Champalimaud, Singularity

Those Silly Fools

"The same way you're laughing at this, they're going to be laughing at us. Not even 100 years from now. 10 years from now: 'Those silly fools didn't know that this bowel disease and this joint disease were the same thing.'"

— TEDMED, Champalimaud, Singularity

Visitation from God

"If a patient dies because of a visitation from God, use code 189."

— Every talk (from the ICD-2 manual, 1909)

Not Otherwise Specified

"You learn a lot about medicine by looking at what we put in this bucket called 'not otherwise specified.' It's all the embarrassing stuff we don't know what to do about."

— NextMed, Champalimaud

Medicine as a Game

"Medicine is practiced synchronously — we write orders, we wait to see what happens, we write more orders. The same way games like chess, checkers, and Go are played. And boy, computers are really great at that."

— LSI

The Triple Question

"Where is this patient going to be in 90 days? What's going to happen in the next year? And what are we going to do about it?"

— TEDxSanFrancisco, TEDxHarkerSchool (his operational definition of precision medicine)

Diabetes Donuts

"We used to call these 'diabetes donuts' and then we realized that would be inappropriate for diabetes. So now we call them life savers."

"And pie would also be inappropriate for diabetes."

— TEDxHarkerSchool, Michigan Omenn

1,500 Ways to Treat Diabetes

"We have 1,500 different ways to start a patient on type 2 diabetes meds. Probably too many. Can we get it down to a thousand? Maybe a hundred? Maybe ten?"

— LSI

Medicine as Computation

If diabetes has 1,500 starting regimens and medicine is practiced synchronously like a game of chess, then the physician's frustration points at a solution: a doctor working with a computer. Butte never framed this as doctors against machines. He framed it as a duty to study every damn thing a health system does, so that every patient benefits from the data the system already holds.

Doctors With Computers

"It's not going to be doctors versus computers. It's going to be doctors with computers versus doctors without computers — who refuse to use computers. That's the real dichotomy."

— Samsung Catalyst

Study Every Damn Thing We Do

"It's not just my responsibility, it should be my duty to study every damn thing we do in our health system, make sure every one of my patients benefits from it."

— LSI

What This Medicine Changes

The conventional reading concluded that precision medicine was a new technology arriving to make care better. That is half right. What Butte changes is the verdict on everything that came before: the imprecise codes, the diabetes donuts, the orders written synchronously and the waiting to see what happens. The real lesson of his medicine is that the imprecision was always the patient's to bear, and the data to fix it was already there.

The Memorial Line

Hiding within these mounds of data is knowledge that could change the life of a patient.

These words belong to a physician who studied every damn thing his health system did, and asked the rest of us to do the same.

ATULISMS | Chapter 6 | On Medicine

Sources

Claim	Source	Link
Atul Butte gave his TEDMED 2012 talk on big data in medicine and precision medicine	TEDMED 2012, "Atul Butte at TEDMED 2012"	tedmed.com
Butte described precisely practicing medicine from a trillion points of data at TEDxSanFrancisco	TEDxSanFrancisco talk recording	youtube.com
Precision medicine is an approach accounting for individual genes, environment, and lifestyle; named in the 2015 Precision Medicine Initiative	NIH / White House Precision Medicine Initiative, 2015	genome.gov
ICD codes lung cancer and diseases under a WHO-standardized classification; historical codes assigned causes like "visitation from God"	World Health Organization, International Classification of Diseases	who.int
Butte led the Bakar Computational Health Sciences Institute at UCSF, converting trillions of data points into diagnostics	UCSF Bakar Computational Health Sciences Institute	bakarinstitute.ucsf.edu

VII. ON AI AND SCALABLE PRIVILEGE

Scalable privilege was Atul Butte's last great phrase, coined after his own 2024 diagnosis, when he realized he could call any famous cancer institute and get information that most people could never reach. He told audiences a mother whose child went undiagnosed by 17 doctors over 3 years typed the symptoms into a chatbot and got an answer for $2.99. The conventional read is that AI in medicine is a frightening, elite technology. The actual rule is that the privilege of the few is now buyable by the many.

The Doors AI Opens

Butte's argument started from a privileged position and ended in a green field. He could open doors most people could not, and he believed AI and informatics would open those doors to everyone — not just in the US, but literally around the world. The same patient-facing decision support that cost a privileged few their connections now costs $9.99 to explain a cancer genome.

AI as Scalable Privilege

"He was in a privileged position. He could call any famous cancer institute and get information on his disease. But most people couldn't open those kind of doors. And he believed that AI and informatics was going to open those doors to everyone around the world, not just in the US, but literally around the world."

— Sam Hogood, Celebration of Life

"It's clunky, it takes privilege to know how to install the OpenAI app. I agree it's privileged. But this is a green field today — patient-facing decision support."

— Danaher

$2.99 to Explain My Echo

"$2.99 to help me explain my echo. $9.99 to explain my cancer genome. Can we really get to that? I think we can."

— Danaher

17 Doctors Couldn't Figure It Out

"This mom — 17 doctors over 3 years could not figure out what their kid had. Just typing the symptoms into GPT — boom, there's a diagnosis."

— Danaher

Demystifying the Machine

If a chatbot can solve what 17 doctors could not, the fear around AI starts to look like mystification. Butte spent his talks pulling back the curtain — there is no wizard, the books cost $17, and the future of computer science is teaching, not gatekeeping.

No Wizard Behind the Curtain

"AI, machine learning — sounds scary. There is no wizard behind the curtain. You can literally buy Dummies books on all of these. $17 each."

— LSI, Danaher

The Future of Computer Science Is Teaching

"A large component of computer science in the future is going to be about being a great teacher. Think about that for a moment."

— TEDxHarkerSchool

What Scalable Privilege Changes

The conventional reading concluded that AI in medicine was a tool for the technically privileged. That is half right. What Butte changes is who holds the doors: the privilege that once required a famous name and a phone call to a cancer institute becomes a $9.99 explanation of a genome, available to a mother in any country. The real shift is not the model — it is who can now open the door.

The Memorial Line

He believed AI would open the doors to everyone around the world.

These are the words of a man who turned his own privilege into a phrase the rest of us could use.

ATULISMS | Chapter 7 | On AI and Scalable Privilege

Sources

Claim	Source	Link
Atul Butte argued AI and informatics would open medical knowledge to everyone around the world	UCSF Bakar Computational Health Sciences Institute profile	bakarinstitute.ucsf.edu
Consumer generative AI (chatbots, GPT models) can produce candidate diagnoses from typed symptoms	OpenAI, GPT model overview	openai.com
Patient-facing AI decision support is an emerging green field in clinical practice	National Library of Medicine, PMC review of clinical AI	ncbi.nlm.nih.gov
Butte gave a TEDMED 2012 talk on big biomedical data	TEDMED 2012	tedmed.com
Butte's career and AI work profiled at UCSF Precision Medicine	UCSF Precision Medicine	precisionmedicine.ucsf.edu

VIII. THE PERSONAL STORIES

Between the data and the drugs, between the TED stages and the grant proposals, there was a man who lost 50 pounds because of a Fitbit, whose own doctor was the last to know. Atul Butte told audiences his genome was 4,000 copies of The Joy of Cooking stacked into two and a half Statues of Liberty, and that he paid $33 for it — less than the parking. The surface story is a scientist telling jokes. The structural story is a teacher disarming an audience so the data could land.

The Body and the Gadget

Butte understood that data is intimidating and humor is not. The first stories he told were about himself: a scale reading 247 pounds after a Hawaii trip, a $40 gadget his billion-dollar health system could not talk to, the Hawthorne effect that improves a number just by measuring it.

The Fitbit and 50 Pounds

"I come back from a trip to Hawaii. I step on a scale — I'm 247 pounds. And I realized, when I hit 250, they can measure me in tons. That's a quarter of a metric ton. What the hell happened to me? Silicon Valley is too easy — I'm driving everywhere."

"Lost 50 pounds in two and a half years."

"It's called the Hawthorne effect — you just measure it, it gets better."

"It's Weight Watchers' rule — you bite it, you write it."

— TEDxSanFrancisco, Singularity, NextMed

The Billion-Dollar System vs. the $40 Gadget

"My doctor's the last person to know I lost all this weight. The most important health intervention ever. Because their billion-dollar system doesn't talk to my $40 gadget."

— TEDxSanFrancisco, Singularity, NextMed

The Genome Made Personal

The same disarming move worked on the genome. Butte made his own DNA a cookbook, a parking-lot purchase, a recipe set to learn rather than fear — and turned medical claims, discharge codes, and maps of death into stories an audience could hold.

Your Genome Is a Cookbook

"Your genome is essentially a cookbook. Your DNA is a lot bigger than The Joy of Cooking — 6 billion base pairs, 3 billion from your mom, 3 billion from your dad. That's about 4,000 copies of The Joy of Cooking. Stack them up: two and a half Statues of Liberty."

— TEDxSF

Don't Be Afraid of Your Genome

"Don't be afraid of your genome. These are your recipes. Learn about them."

— TEDxSF

The Gene for Compliance

"We have yet to find the gene and the genome for compliance with medical care. I kind of think Steve Quake doesn't have that gene."

— TEDxSF (on his patient refusing statins)

The Daughter at TEDMED

"When I told my 9-year-old daughter Kimmy that I was giving a TEDMED talk, she was thrilled. And it's kind of amazing to me that a 9-year-old already knew what a TED Talk was."

— TEDMED 2012

Maps of Death and Disease

"Think of Google Maps — Google Maps takes you to pleasant destinations. I'm trying to figure out how you're gonna die. Perhaps the exact opposite of Google Maps."

— TEDxSanFrancisco, Michigan Omenn

Parking Lot Genomes

"$33 for your genome. I think some of you are probably going to end up paying more than that for parking today."

— TEDxSF, Singularity

The Restaurant Bill vs. the Menu

"Does it really help you to look at the bill? Wouldn't you rather see what we ordered? And that's the difference between medical claims and medical records."

— TEDxSanFrancisco

The Oak Ridge Boys

"Discharge disposition is like that Oak Ridge Boys song — you don't have to go home but you can't stay here."

— TEDxSanFrancisco

Kittens Playing with Pianos

"Most of these zettabytes of data are these incredible videos of kittens on YouTube. Entertainment value. Perhaps no scientific value. But there is scientific data in the zettabytes as well."

— Champalimaud, Singularity, Michigan Omenn

What the Personal Stories Change

The conventional reading concluded that these were just the warm-up jokes before the real science. That is half right. What the personal stories change is the order of trust: the man who lost 50 pounds on a Fitbit, who paid $33 for his genome, who admitted his doctor was the last to know, earned the audience before the data started. The real lesson is that he made the science human by making himself the patient.

The Memorial Line

Even data danced when he looked at it with joy.

These are the stories a scientist told about himself so that the rest of us would trust the data he carried.

ATULISMS | Chapter 8 | The Personal Stories

Sources

Claim	Source	Link
The Hawthorne effect describes behavior improving simply because it is being measured	Encyclopaedia Britannica, Hawthorne effect	britannica.com
A human genome contains about 6 billion base pairs, 3 billion inherited from each parent	National Human Genome Research Institute	genome.gov
Atul Butte described his genome as a cookbook in his TEDxSanFrancisco talk on a trillion points of data	TEDxSanFrancisco	youtube.com
Consumer wearables like Fitbit track weight and activity outside the clinical record	Fitbit, official site	fitbit.com
Butte sized his genome at 4,000 copies of The Joy of Cooking stacked into two and a half Statues of Liberty	The Joy of Cooking, official	thejoykitchen.com

IX. ON MENTORSHIP AND CAREER

Chris Longhurst proposed studying medication errors across a children's hospital, and Atul Butte told him to think bigger — twice — until the project became all-cause mortality. Across the Michigan Omenn lecture and the 2024 Collen Award, Butte ran one operating system for every student who passed through his lab: credit is infinitely divisible, junior faculty is a state of funding not a state of mind, and your peers are all over the world. The conventional read is that mentorship is about protecting your own credit. The actual rule is that you give it all away and there is always plenty left.

The Operating System

Butte's mentorship was a set of instructions he repeated until they became reflex. Think bigger. Set your peers high. You determine your future, not NIH. Solve real-world problems. Cheer your colleagues. Treat failure as a badge. Every line was an order to a student to aim past what they thought possible.

Think Bigger

Chris Longhurst: "I'm going to look at medication errors across the whole children's hospital." Atul: "Chris, like everybody's done that. You got to think bigger." Chris: "We could probably impact bronchiolitis outcomes." Atul: "No, Chris, you got to think bigger." (They ended up looking at all-cause mortality.)

— Celebration of Life

Credit Is Infinitely Divisible

"He'd have that sign that said, 'Credit is infinitely divisible,' because you can put as many co-authors on the paper who want to lick that cookie."

— Chris Longhurst, Celebration of Life

Licking the Cookie

"He started talking to me about 'licking the cookie.' In his world, Atul had tons of cookies and he wanted everyone to lick the cookie. He did it with a joyfulness."

— David Rubin, Celebration of Life

Junior Faculty as a State of Funding

"Junior faculty should be a state of funding, not a state of mind. Junior faculty should be trying to kick their mentors' asses."

— Michigan Omenn

Set Your Peers High

"Set the level of your peers as high as you can. Your peers are all over the world, in countries you've never been to. That's your new level."

— Michigan Omenn

You Determine Your Future

"You determine your future, not NIH."

— Michigan Omenn

Solve Real-World Problems

"I'm an h-index junkie. I'm vain. I'll admit it. Citations are fun games. But they're just games. Solve real-world problems."

— Michigan Omenn

Cheer Your Colleagues

"We have to cheer our colleagues more to get more total funding into science, not keep putting each other down."

— Michigan Omenn

Failure Is a Badge

"That's great, Chris, because you've learned what you want to do and you've learned that you can be a startup failure. And that's a badge of pride in Silicon Valley."

— Chris Longhurst, Celebration of Life

The Joy in the Work

Late in his career, the same mentor who told students to think bigger spoke about the treasure of every working day and the introvert who still loved to write software code. The students he had aimed past their limits came back, after, to say what they had learned.

Every Day Is a Treasure

"I feel like every day that I get to go to work and to talk to people is a treasure. I don't take any day for granted right now."

— Collen Award 2024

The Introvert Who Loves Code

"I love my work. I love talking to people, mentoring folks, but still the little introvert in me still loves that I get to write software code."

— Collen Award 2024

What They Said on LinkedIn (After)

"Nearly everything that I know about doing science I learned from Atul. Warp speed, Atul."

— Matthew Kan, MD, PhD. Butte Lab postdoc 2016-2017. Assistant Professor of Pediatrics, UCSF. Source: LSI Cover Story, lifesciencemarketresearch.com, 2025.

When Kan told Atul about his faculty appointment and newborn son, Atul replied: "...congrats on your new Large Language Model, and so great to see exponential growth in those neurons!!" Source: ibid.

"He taught me how to ask questions, how to think big, how to mentor and bring collaborative teams together. He was a true innovator and a champion."

— Marina Sirota, PhD. Acting Director, Bakar Computational Health Sciences Institute, UCSF. Source: LSI Cover Story, lifesciencemarketresearch.com, 2025.

"He didn't seek credit; he invited others in."

— Joe Smith, MD, PhD. Senior VP & Chief Scientific Officer, BD. Source: LSI Cover Story, lifesciencemarketresearch.com, 2025.

"He had the entire audience spellbound. Not just learning, but laughing, inspired, and moved to act. You didn't just want to hear him speak — you wanted to be part of his orbit."

— Lisa Carmel, Chair of Business Development, Mayo Clinic. Source: LSI Cover Story, lifesciencemarketresearch.com, 2025.

"He was a big fan of Tufte. He sent us to an expensive Tufte workshop."

— Atul Butte Mafia WhatsApp group (120363402783687844@g.us), September 27, 2025. Source: WhatsApp ChatStorage.sqlite, compiled SOP-009.

What This Mentorship Changes

The conventional reading concluded that academic success is a zero-sum fight for credit and funding. That is half right. What Butte changes is the arithmetic: credit is infinitely divisible, peers are set high across the whole world, and a startup failure is a badge of pride. The real lesson his students carried is that you bring collaborative teams together by inviting people in, not by guarding the cookie.

The Memorial Line

Nearly everything that I know about doing science I learned from Atul.

These are the words of the students who came back to say what one mentor's operating system made of them.

ATULISMS | Chapter 9 | On Mentorship and Career

Sources

Claim	Source	Link
Atul Butte mentored students to think bigger and bring collaborative teams together at UCSF	UCSF Bakar Computational Health Sciences Institute	bakarinstitute.ucsf.edu
Chris Longhurst is a physician-informatician who trained in Butte's orbit	UC San Diego Health, Chris Longhurst profile	health.ucsd.edu
Marina Sirota directs the Bakar Computational Health Sciences Institute at UCSF	UCSF Bakar Institute, Marina Sirota	sirotalab.ucsf.edu
The h-index measures researcher productivity and citation impact	Hirsch, "An index to quantify an individual's scientific research output," PNAS 2005	pnas.org
At the Collen Award, Butte called every working day a treasure and praised the introvert who still loved to write software code	American College of Medical Informatics, Morris F. Collen Award	amia.org

X. THE MAN THEY REMEMBERED

On June 21, 2025, eight days after Atul Butte died, the room at the Celebration of Life held 500 people and was not enough. Peter Embi, Zach Kohane, Ken Mandl, Chris Longhurst, Sam Hawgood, Marina Sirota, and Gini Deshpande each told a different facet of the same man — the zebra unicorn, the screen-scraping genius, the man from Turkey whose life changed in one meeting. The surface story is a memorial service for a brilliant scientist. The structural story is a community discovering it had inherited an operating system for how to treat people.

The Words Spoken That Day

Each speaker told a different facet of the same man. The zebra who became the rarest zebra. The colleague who simply showed up. The mentor who answered an email a month later from a keynote in Morocco. The husband who proposed in PostScript. These are the words spoken at the Celebration of Life, from the people who knew him best.

The Zebra Unicorn

"In medicine, we teach: when you hear hoofbeats, think horses, not zebras. The thing is zebras exist and Atul knew that. He spent his career working to help the zebras. And then ironically he became a zebra."

"He wasn't satisfied just being special. He couldn't just be a zebra. He needed to be the rarest of zebras. Of course. So he had one of the rarer forms of one of the rarest cancers. He was a zebra unicorn."

— Peter Embi, Celebration of Life

He Showed Up

"Here's something about Atul that you can all emulate. And you don't need his off-the-charts IQ. You don't need his one-of-a-kind smile. And you don't need his boundless energy. Atul showed up."

— Ken Mandl, Celebration of Life

The Man from Turkey

"A gentleman in Turkey: 'I only met Atul once. It was at a keynote in Morocco. I was so inspired and he followed up a month later with an email to make sure I was working on those ideas.' He said, 'Atul completely changed the course of my life, and I only met him that one time.'"

— Chris Longhurst, Celebration of Life

Scalable Privilege (the origin)

"He used his own diagnosis, his own disease, to introduce the term 'scalable privilege.' He could call any famous cancer institute and get information. But most people couldn't open those doors. And he believed AI was going to open those doors to everyone around the world."

— Sam Hogood, Celebration of Life

"He had an absolute great talent at framing things in ways that were catchy. 'Scalable privilege.' Come on. So good."

— Zach Kohane, Celebration of Life

Never Too Busy

"Despite the incredible demands on his time, Atul was never too busy — and I mean that, never too busy — to mentor a junior colleague, to celebrate a friend's success, or to share a thought about the absurdities of academic life."

— Philip Payne, Celebration of Life

The Screen-Scraping Genius

"He created a program that screen-scraped the Cerner lab user interface and wrote in this arcane language, PostScript, a program that generated very tight summaries of patient problem lists. The best thing I'd ever seen, ever. We were all scared that Atul would stop doing it, and sure enough when he left we never had it again."

— Zach Kohane, Celebration of Life

Fight to the End

"He would be in every meeting — first in person, then on Zoom, then on Zoom off camera, then on Zoom without audio, but in the chat. Right to the very end."

— Sam Hogood, Celebration of Life

"He was on emails and texts with me about institute matters until only a month or two ago, even from the ICU."

— Marina Sirota, Celebration of Life

Cheers, Another Morning

"After his diagnosis, we started a new ritual. We would clink our mugs and say, 'Cheers, another morning.' Even from the hospital, he insisted on continuing this ritual, determined to savor every moment we had together."

"So, Atul, wherever you are today, cheers. Here's to another morning."

— Gini Deshpande, Celebration of Life

Five Seconds

"Friends who've known me for decades were stunned when I told them that just 5 seconds after meeting him, I knew I was going to marry this man."

— Gini Deshpande, Celebration of Life

The PostScript Proposal

"He proposed to Gini by taking one of her manuscripts, edited it in PostScript, changed a part to say 'Will you marry me?' and came over to Gini and said, 'Look, they got this wrong in your paper.' And she got all ready to be outraged."

— Jesse Tenenbaum, Celebration of Life

The Spark

"Ideas lit the room. He spoke and silence listened. Fire glowed within his eyes. He filled the air and hearts, left them brighter. Even data danced when he looked at it with joy."

— Eve Lucier, Celebration of Life

On Problems That Matter

"In a world that often moves too fast, what Atul reminded us was that we should spend our time on problems that matter with people who matter."

— Philip Payne, Celebration of Life

His Name

"Atul's name carries deep meaning. It means incomparable, one of a kind, unique. And in every aspect of his life, Atul truly lived up to that name."

— Gini Deshpande, Celebration of Life

The Clan Remembers

Beyond the speakers at the Celebration of Life, the students he called his Mafia kept remembering — in the WhatsApp group created the day after he died, in the LinkedIn tributes, in the babies born that December. The man who showed up for everyone was now the man everyone showed up for.

What the Student Said (Mafia)

"Of all my mentors, Atul was easily the nicest. And biggest balls! The day he met me he asked me if I was crazy to be in residency with 3 children at Stanford. I soon dropped out and joined his lab. He taught me the secret to happiness is to NOT practice."

"Meeting with Atul was among the most exhilarating experiences ever for me. Not that I've ever done cocaine, but I'd imagine it's like how I felt after meeting 1:1 with him."

"I'd take Atul as a god father."

— Dexter Hadley. Source: Atul Butte Mafia WhatsApp (120363402783687844@g.us), June 14-15, 2025. WhatsApp ChatStorage.sqlite, compiled SOP-009.

What They Said on LinkedIn (After)

"The passing of Atul Butte still weighs heavy on me. There are some people that are just special. I'm going to miss the insightful conversations. Thanks for being an inspiration Atul."

— DJ Patil, former U.S. Chief Data Scientist. Source: LSI Cover Story, lifesciencemarketresearch.com, 2025.

"He genuinely loved seeing people succeed."

"He lived vicariously through me and all the crazy entrepreneurial things that I did."

— Tarangini (Gini) Deshpande, PhD. Wife. Cancer biologist. Named the "Atul Butte Mafia." Source: UCSF obituary, ucsf.edu/news/2025/06/430236, June 2025.

What They Said Three Months Later (Mafia)

"Three months since Atul, but still feels like just yesterday."

— Atul Butte Mafia WhatsApp (120363402783687844@g.us), September 13, 2025. Source: WhatsApp ChatStorage.sqlite, compiled SOP-009.

"Our friendships are another thing we owe to Atul."

— Atul Butte Mafia WhatsApp (120363402783687844@g.us), September 1, 2025. After a gathering at Sanchita and Suman's. Source: ibid.

"Ideas lit the room. He spoke and silence listened. Fire glowed within his eyes. He filled the air and hearts, left them brighter. Even data danced when he looked at it with joy."

— Yves Lussier, MD, Chair of Biomedical Informatics, University of Utah. Source: LSI Cover Story, lifesciencemarketresearch.com, 2025.

Butte Lab Babies

"Butte lab babies!!!"

— Atul Butte Mafia WhatsApp (120363402783687844@g.us), December 30, 2025. Three babies born that month — Ben Glicksberg's, Dexter's, and Matt's. Dexter replied: "Takes three points to make a curve." Source: WhatsApp ChatStorage.sqlite, compiled SOP-009.

What These Words Change

The conventional reading concluded that a memorial is a backward-looking ritual that ends with the burial. That is half right. What the man they remembered changes is the direction of the memory: the operating system he ran — show up, think bigger, invite people in — propagated to a clan that now runs it without him. The real lesson of the Celebration of Life is that the words outlived the man, exactly as his signature lines always had.

The Memorial Line

Cheers, another morning.

These are the words a community spoke over the man who taught them all to show up.

ATULISMS | Chapter 10 | The Man They Remembered

Sources

Claim	Source	Link
Atul Butte's Celebration of Life was livestreamed in June 2025 after his death	UCSF obituary for Atul Butte	ucsf.edu
Butte died of a malignant peripheral nerve sheath tumor, a rare cancer	National Cancer Institute, MPNST overview	cancer.gov
PostScript is a page-description programming language Butte used to screen-scrape and edit documents	Adobe, PostScript language	adobe.com
Cerner is an electronic health record system whose lab interface Butte screen-scraped	Oracle Health (Cerner)	oracle.com
Marina Sirota leads the Bakar Computational Health Sciences Institute at UCSF	UCSF Bakar Institute, Sirota Lab	sirotalab.ucsf.edu

XI. THE NUMBERS

Atul Butte loved statistics, so this chapter gives him his own. The memorial corpus runs to 48 transcripts, 50 videos, 18 hours of camera time, 958 pages, and 239,612 words spanning 2004 to 2025. He said "patient" 706 times, "big data" 142 times, "scalable privilege" 8 times, and "zebra unicorn" twice. The count went up. The rule got smaller.

The Corpus by the Numbers

The numbers are the man made measurable. Forty-eight transcripts processed into 1,198,061 characters; the signature phrases ranked by frequency; the eras from Stanford to UCSF to Memorial; the career from 1969 to 2025, age 55, ended by MPNST.

The Tally

48 transcripts processed
50 videos cataloged
18+ hours of Atul on camera
958 pages of text
239,612 words
1,198,061 characters

SIGNATURE PHRASES (frequency across all transcripts):
├── "patient"              706 mentions
├── "big data"             142
├── "clinical trial"       137
├── "precision medicine"    88
├── "machine learning"      64
├── "data science"          47
├── "trillion points"       31
├── "garage biotech"        12
├── "scalable privilege"     8
├── "frozen knowledge"       4
└── "zebra unicorn"          2

TALKS SPANNING: 2004-2025
ERAS: Stanford → UCSF → Memorial
CAREER: 1969-2025. Age 55. MPNST.

What the Numbers Change

The conventional reading concluded that a word-frequency tally is a curiosity, a footnote to the talks. That is half right. What the numbers change is the proof: 706 mentions of "patient" against 8 of "scalable privilege" show the man returned, talk after talk, to the people rather than the phrase. The real lesson in the tally is that the repetition was instruction, and the frequencies are the fingerprint of a 21-year argument.

The Memorial Line

239,612 words, and the one he said most was patient.

These are the numbers of a man who counted everything so that every patient could be counted.

ATULISMS | Chapter 11 | The Numbers

Sources

Claim	Source	Link
Atul Butte's career spanned 1969 to 2025 across Stanford, UCSF, and Memorial-era roles	UCSF obituary for Atul Butte	ucsf.edu
Butte built tools converting trillions of molecular and clinical data points into insight	UCSF Bakar Computational Health Sciences Institute	bakarinstitute.ucsf.edu
Big data describes datasets too large for traditional processing, measured in characters and words	Encyclopaedia Britannica, big data	britannica.com
MPNST (malignant peripheral nerve sheath tumor) is the rare cancer that ended Butte's career at age 55	National Cancer Institute	cancer.gov
Word-frequency analysis of a transcript corpus reveals an author's recurring signature phrases	Stanford NLP, text analysis overview	nlp.stanford.edu

XII. THE ARC

The arc is short: Atul Butte, 1969 to 2025, 55 years. Harvard screen-scraping in the 1990s, Stanford and "Microarrays for an Integrative Genomics" in 2004, the TEDMED talk in 2012 that defined him, the Bakar Institute at UCSF in 2015, the Collen Award in 2024, death by MPNST in June 2025, and the Celebration of Life on June 21. The story everyone tells is a career that ended in 2025. The story nobody tells is a memorial that began in 2026.

The Timeline

The arc moves through institutions and phrases at once. Each stop on the timeline is a place and a sentence: Harvard and the PostScript proposal, Stanford and integrative genomics, TEDMED and "data is frozen knowledge," UCSF and "a trillion points of data," the Collen Award and "every day is a treasure," and finally ATULISMS in 2026 — commissioned by Gini, written by Dexter.

The Stops

Atul Butte (1969-2025)
    │
    ├── Harvard (1990s) ──── Screen-scraping Cerner, PostScript proposals
    │
    ├── Stanford (2004) ──── "Microarrays for an Integrative Genomics"
    │
    ├── TEDMED (2012) ──── "Data is frozen knowledge"
    │                       The talk that defined him
    │
    ├── UCSF (2015) ──── Bakar Institute
    │                     "A trillion points of data"
    │
    ├── Collen Award (2024) ──── "Every day is a treasure"
    │
    ├── Dies (Jun 13, 2025) ──── MPNST, age 55
    │                             Zebra unicorn
    │
    ├── Celebration of Life (Jun 21) ──── "Cheers, another morning"
    │
    └── ATULISMS (2026) ──── The words. The man. The memorial.
                              Commissioned by Gini. Written by Dexter.

What the Arc Changes

The conventional reading concluded that a life arc ends at death — Harvard to Stanford to UCSF to the grave. That is half right. What the arc changes is its own endpoint: the timeline does not stop at June 2025 but continues into ATULISMS in 2026, the words organized by theme and the man made legible through his own voice. The real lesson of the arc is that the data is still frozen, and someone will melt it.

The Memorial Line

The arc does not end here.

This book is the words, the man, and the memorial, commissioned by Gini and written by Dexter.

ATULISMS | Chapter 12 | The Arc

Sources

Claim	Source	Link
Atul Butte's career ran from 1969 to 2025, ending at age 55	UCSF obituary for Atul Butte	ucsf.edu
Butte's 2004 work "Microarrays for an Integrative Genomics" framed his data-driven genomics career	MIT Press, Microarrays for an Integrative Genomics	mitpress.mit.edu
Butte gave a defining TEDMED talk in 2012	TEDMED 2012	tedmed.com
Butte founded and led the Bakar Computational Health Sciences Institute at UCSF in 2015	UCSF Bakar Institute	bakarinstitute.ucsf.edu
MPNST is the rare cancer that ended Butte's arc in June 2025	National Cancer Institute	cancer.gov

XIII. THE CLAN

The Atul Butte Mafia WhatsApp group was created on June 14, 2025 — the day after Atul died — and reached 66 members and 485 messages. Gini Deshpande named it; the clan filled it; Dexter Hadley called it an honor to be part of Atul's legacy. The story everyone tells is that a community gathers to grieve and then disperses. The story nobody tells is that this one kept building, treating a man's death as the start of his thesis rather than its end.

The Group That Grew

The clan did more than mourn. It hosted memorials, posted tributes, and turned grief into the work Atul championed — drug repurposing, N-of-1 treatment, rare-cancer initiatives in his name. From the day the group was created through the babies born that December, the Mafia kept Atul's operating system running.

The Name

"In the early days of NuMedii, someone came up to me and asked if Atul was involved. When I said yes, he stated 'oh, you must be part of the Atul Butte Mafia'. Atul found it absolutely hilarious."

— Tarangini (Gini) Deshpande, PhD. Source: Atul Butte Mafia WhatsApp (120363402783687844@g.us), June 23, 2025. WhatsApp ChatStorage.sqlite, compiled SOP-009.

Day One

The WhatsApp group was created June 14, 2025 — the day after Atul died.

31 members by nightfall. 66 by the end.

"Celebrating 10 years of the Butte Lab."

— Joel. Shared a video. Source: Atul Butte Mafia WhatsApp, June 14, 2025. WhatsApp ChatStorage.sqlite, compiled SOP-009.

"Honored to be a member of 'Atul's clan' and thus his legacy."

— Dexter Hadley. Source: Atul Butte Mafia WhatsApp, June 14, 2025. WhatsApp ChatStorage.sqlite, compiled SOP-009.

The Gathering

Marina hosted a gathering at her home. Rong Chen posted a LinkedIn memorial. Bin Chen wrote his — "seven years ago." Beau Norgeot wrote a three-part tribute.

Source: Atul Butte Mafia WhatsApp, June 15, 2025. WhatsApp ChatStorage.sqlite, compiled SOP-009. LinkedIn posts: linkedin.com/posts/beau-norgeot-phd (Parts 1-3).

"He taught me the secret to happiness is to NOT practice."

— Dexter Hadley, June 15, 2025. Source: ibid.

The Funeral and Celebration

Funeral: Friday, June 20, 2025. Skylawn Memorial Park, San Mateo. Celebration of Life: Saturday, June 21, 2025. Livestreamed.

youtu.be/YZgPZcRCtig

Donations directed to: UCSF Foundation, MPNST Research Fund.

Source: Atul Butte Mafia WhatsApp, June 17-21, 2025. WhatsApp ChatStorage.sqlite + Skylawn tribute wall, skylawnmemorialpark.com/obituaries/atul-butte.

The Rare Cancer

Gini: "You should involve Gini" — members proposed a rare cancer initiative in Atul's name.

Sam, Emily, Krishna, Alan — cancer center leadership at UCSF — all named in the discussion.

Source: Atul Butte Mafia WhatsApp, June 15-19, 2025. WhatsApp ChatStorage.sqlite, compiled SOP-009.

The Birthday Memorial

August 26, 2025. Atul's birthday. Sanchita and Suman co-hosted, with Gini.

"Our friendships are another thing we owe to Atul."

— September 1, 2025. After the gathering. Source: Atul Butte Mafia WhatsApp. WhatsApp ChatStorage.sqlite, compiled SOP-009.

Three Months

"Three months since Atul, but still feels like just yesterday."

— September 13, 2025. Source: Atul Butte Mafia WhatsApp. WhatsApp ChatStorage.sqlite, compiled SOP-009.

The Thesis Made Real

The clan's grief became Atul's argument in practice. An FDA clearance for a drug-repurposing diagnostic, talks dedicated to him at MIT, MSK, and NYU, a member treating an ultrarare pediatric disorder with a repurposed drug — each was the work he championed, carried forward by the people he gathered.

The Legacy in Motion

September 30, 2025 — Inflammatix receives FDA clearance for a sepsis diagnostic. Drug repurposing from public data. The work Atul championed, cleared by the FDA.

"Just wish Atul was around to see this."

Source: Atul Butte Mafia WhatsApp, September 30, 2025. WhatsApp ChatStorage.sqlite, compiled SOP-009.

October 15, 2025 — MIT talk dedicated to Atul. October 16, 2025 — MSK and NYU talks dedicated to Atul.

Source: Atul Butte Mafia WhatsApp, October 2025. WhatsApp ChatStorage.sqlite, compiled SOP-009.

Gini's Call

"Looking for ideas... create a project in Atul's memory."

— Gini Deshpande, December 9, 2025. Precision medicine. N-of-1 initiative.

"I just successfully treated a pediatric patient with an ultrarare genetic disorder by repurposing an FDA approved drug. This specific idea I came up with while in Atul's lab."

— A member replied. The thesis made real.

Source: Atul Butte Mafia WhatsApp, December 9, 2025. WhatsApp ChatStorage.sqlite, compiled SOP-009.

Butte Lab Babies

December 30, 2025. Three babies born that month — Ben Glicksberg, Dexter Hadley, and Matt.

"Butte lab babies!!!"

"Takes three points to make a curve."

— Dexter Hadley.

Source: Atul Butte Mafia WhatsApp, December 30, 2025. WhatsApp ChatStorage.sqlite, compiled SOP-009.

The Roll Call

66 members. 485 messages. Created the day after he died.

Resolved: Ben Glicksberg. Boris Oskotsky. Brenda Miao. Kelly Zalocusky. Andrew. Greg. Harry. Karthik. Dexter Hadley.

Named in messages: Marina Sirota. Gini Deshpande. Beau Norgeot. Rong Chen. Bin Chen. Travis. Joel. Zak. Nova. Kimi. Sam. Emily. Sanchita. Suman. Syzygy (Atul's child — discovered a galaxy). Rahul. Michael Januszyk. Nadav. Idit. Matt. Peter Embi. Peter Szolovits. Krishna. Alan.

51 unresolved. WhatsApp encryption. They're in there. We just can't read their names.

Source: WhatsApp ChatStorage.sqlite, group 120363402783687844@g.us. Compiled SOP-009, 2026-02-06. Full alumni list: buttelab.ucsf.edu/people/.

The Competition

PMWC 2026. March 4-6, Santa Clara.

Track: "Atul Butte Company Competition." $20,000 prize.

The competition is named for him. One of his students is competing.

What the Clan Changes

The conventional reading concluded that a memorial group is a temporary place to grieve. That is half right. What the clan changes is the verb: 66 members did not just remember Atul, they continued him — repurposing drugs, treating rare disorders, dedicating talks, naming a competition. The real lesson of the clan is that the thesis Atul preached for two decades became real in the hands of the people he gathered the day after he died.

The Memorial Line

Our friendships are another thing we owe to Atul.

This chapter is the record of a clan that turned a man's death into the start of his thesis.

ATULISMS | Chapter 13 | The Clan

Sources

Claim	Source	Link
The Atul Butte Mafia formed among his lab alumni and collaborators after his death	UCSF obituary for Atul Butte	ucsf.edu
Inflammatix develops sepsis diagnostics built on drug-repurposing and gene-signature science	Inflammatix, official site	inflammatix.com
Drug repurposing finds new uses for FDA-approved drugs from existing data	National Center for Advancing Translational Sciences, drug repurposing	ncats.nih.gov
NuMedii was the drug-discovery company built on Butte-lab public-data methods	NuMedii, company background	crunchbase.com
The Precision Medicine World Conference hosts company competitions in Santa Clara	Precision Medicine World Conference	pmwcintl.com

Commissioned by Tarangini (Gini) Deshpande, PhD. "In the early days of NuMedii, someone came up to me and asked if Atul was involved."

TRANSITIONS — Connective Prose

inherits: hadleylab-canonic/CONTENT/BOOKS/ATULISMS/CHAPTERS

These transitions are Marina Sirota's voice, narrating her arc through Atul Butte's world from September 2006 to the institute she now directs. Each connects one quote chapter to the next — from the signature lines, to data, to open science, to drug repurposing, to medicine, to the man himself. The surface story is connective tissue between quote chapters. The structural story is one student's career retracing a mentor's argument across twenty years.

These transitions connect the quote chapters. They are Marina's voice.

Before Chapter 1: The Signature Lines

Some phrases outlive the people who said them. Atul's did.

He had maybe six sentences that he repeated in every talk, across every audience, for twenty years. The words did not change. The delivery did not change. The conviction did not change. If you heard Atul speak once, you heard "data is frozen knowledge." If you heard him speak twice, you realized he meant it literally. If you heard him speak a dozen times — as Marina did, as everyone in his orbit did — you realized the repetition was not habit. It was instruction.

These are the signature lines. The phrases that 706 mentions of "patient" and 142 mentions of "big data" orbit around. The rhetorical DNA.

Before Chapter 2: On Data

Marina's PhD thesis was built on Atul's data philosophy: that publicly available gene expression data — millions of samples sitting untouched on servers — could reveal new uses for existing drugs. She did not generate new data. She used data that other labs had already published. Atul's argument was that this was not only sufficient but better: "wisdom of the crowd," he called it. You do not trust any one experiment. You trust what they show you in common.

What follows is what he said about data — across two decades, to audiences from Singularity University to TEDx to the Collen Award. The message never drifted. Data is not oil. Data is soil. You plant ideas, and data helps them grow.

Before Chapter 3: On Open Science

When Marina was building the drug repurposing pipeline in Atul's lab at Stanford, she was not generating proprietary data. She was downloading it. Free. From GEO. From public repositories that the NIH had funded and that most researchers ignored.

Atul found this hilarious. He would tell audiences: "Let me pretend I'm writing a brand-new NIH grant. I propose that I can get 3,700 of the best breast cancer labs in the world to share data with me for free. You're already laughing." The audience always laughed. The point always landed: the data was already there, waiting.

This chapter is his case for open science — the argument that the most expensive data in America was sitting unused, that the priesthood of elites was dissolving, and that a high school kid with a laptop had more samples available than any breast cancer researcher would ever have in their own lab.

Before Chapter 4: On Drug Repurposing

This is where Marina's career began. Her thesis in Atul's lab at Stanford — published in Science Translational Medicine — demonstrated that you could take a disease's gene expression signature, flip it, and find existing drugs that reverse it. "Match.com for drugs," the reporters called it. Opposites attract.

The work became NuMedii, the company Gini Deshpande led. From public data to acquisition in twenty-four months. The pipeline that Marina started as a graduate student now runs at single-cell resolution, targeting Alzheimer's, published in Cell in 2025. The foundation Atul laid — turn the crank, find the match, test the mouse — is still the architecture.

Before Chapter 5: On Garage Biotech

Atul believed the next Amgen would start in a garage. He said it at Singularity. He said it at Champalimaud. He said it at Michigan. He meant it every time.

The argument was straightforward: the tools had been democratized. You could buy a mouse model off the internet with a credit card. You could download more breast cancer samples than any lab would ever generate. You could screen drugs computationally for the cost of a cloud subscription. The priesthood was over. The garage was open.

Marina saw this play out. Her thesis was garage biotech before the term existed — computational drug discovery using publicly available data, no wet lab required until the mouse validation stage. The economics Atul preached were the economics she lived.

Before Chapter 6: On Medicine

Atul was a physician. This is sometimes forgotten behind the data science, the drug repurposing, the TED talks, the startup acquisitions. But he trained at Harvard. He did residency. He saw patients. And his frustration with medicine — its imprecision, its coding absurdities, its 1,500 ways to start a patient on diabetes meds — came from direct clinical experience.

What follows is the physician speaking. Not the data scientist. Not the entrepreneur. The doctor who looked at ICD codes and saw "visitation from God" as a cause of death. The doctor who knew that precision medicine was an admission that everything before it was imprecision.

Before Chapter 7: On AI and Scalable Privilege

"Scalable privilege" was Atul's last great phrase. He coined it after his own diagnosis — when he realized that he could call any famous cancer institute and get information on his disease, but most people could not open those doors. AI, he believed, would open them. For everyone. Not just in the US, but around the world.

Marina's current work — extending the drug repurposing pipeline to Alzheimer's using clinical data from 8 million UCSF patients — is scalable privilege in practice. The infrastructure Atul built at Bakar (the Information Commons, the UC Health data system) is what makes it possible. The privilege of having 8 million deidentified records is being scaled to every researcher who can access the system.

Before Chapter 8: The Personal Stories

Between the data and the drugs, between the TED stages and the grant proposals, there was a man who lost 50 pounds because of a Fitbit, whose doctor was the last to know, who told audiences his genome was 4,000 copies of The Joy of Cooking stacked into two and a half Statues of Liberty.

These are the stories he told about himself. The disarming ones. The ones that made audiences trust him before the science started. Atul understood that data is intimidating and humor is not.

Before Chapter 9: On Mentorship and Career

"Think bigger."

Marina heard this. Chris Longhurst heard this. Everyone who ever proposed a project to Atul heard this. The scope was never enough. The ambition was never enough. Junior faculty should be trying to kick their mentors' asses. Your peers are all over the world. Solve real-world problems. Cheer your colleagues.

This chapter is the mentorship manual — the operating system he ran for every student, postdoc, and junior faculty member who passed through his lab. Credit is infinitely divisible. Give it away every chance you get. There is always plenty left for you.

Before Chapter 10: The Man They Remembered

On June 21, 2025, eight days after Atul died, the room at the Celebration of Life held five hundred people and was not enough. What follows are the words spoken that day — by Peter Embi, Zach Kohane, Ken Mandl, Chris Longhurst, David Rubin, Jesse Tenenbaum, Sam Hawgood, Marina, Gini, Eve Lucier, Philip Payne.

Each speaker told a different facet of the same man. The zebra unicorn. The screen-scraping genius. The man from Turkey whose life changed in one meeting. The PostScript proposal. The fight to the end — on Zoom off camera, then without audio, then in the chat.

Before Chapter 11: The Numbers

The real lesson of the tally is the fingerprint of an argument repeated for twenty years. Atul loved statistics. Marina said it first: "We all love data and statistics." So here are his.

Forty-eight transcripts. Fifty videos. Eighteen hours on camera. 958 pages. 239,612 words. Seven hundred and six mentions of "patient." One hundred and forty-two of "big data." Thirty-one of "trillion points." Four of "frozen knowledge." Two of "zebra unicorn."

Talks spanning 2004 to 2025. Three eras. One career. Age 55.

Before Chapter 12: The Arc

The arc is short. 1969 to 2025. Fifty-five years.

Harvard. Stanford. UCSF. TEDMED. Collen Award. Diagnosis. Death. Memorial. And then this book — the words, organized by theme, the man made legible through his own voice.

The arc does not end here. Marina continues the science. Gini carries the legacy. The Mafia carries the memory. The data is still frozen. Someone will melt it.

These transitions carry Marina's voice from one chapter to the next, retracing a mentor's argument across twenty years.

TRANSITIONS | ATULISMS | CONNECTIVE PROSE

Sources

Claim	Source	Link
Marina Sirota directs the Bakar Computational Health Sciences Institute at UCSF	UCSF Bakar Institute, Sirota Lab	sirotalab.ucsf.edu
Public gene-expression data from repositories like GEO enables computational drug repurposing	National Center for Biotechnology Information, Gene Expression Omnibus	ncbi.nlm.nih.gov
Drug repurposing reverses a disease's gene-expression signature with existing drugs	National Center for Advancing Translational Sciences	ncats.nih.gov
Atul Butte's defining TEDMED 2012 talk argued data is frozen knowledge waiting to flow	TEDMED 2012	tedmed.com
The wisdom of the crowd describes aggregating many independent observations for better answers	Encyclopaedia Britannica, wisdom of the crowd	britannica.com

Abstract

Body

Table of Contents

PREFACE

Abstract

Foreword

Preface

Compiler's Note

The Ledger

Sources

I. THE SIGNATURE LINES

The Signature Phrases

"Data is frozen knowledge"

"A trillion points of data"

"If you want to change the world, you can't just keep writing papers about it"

"It will be a national tragedy"

What the Phrases Meant

"Everything is connected"

"One word: hope"

Sources

II. ON DATA

Data Is Soil, Not Oil

Data Is Not Oil — Data Is Soil

The Most Expensive Data in America

The Question Is the Hard Part

Data By Itself Doesn't Do Anything

Big Data Comes from Small Packages

Sources

III. ON OPEN SCIENCE

The Open-Science Argument

Retroactive Crowdsourcing

If a High School Kid Can Do It

The Ridiculous NIH Grant

I'm Giving Away the Secrets

The Priesthood

The Taxpayer Chain

Sources

IV. ON DRUG REPURPOSING

Finding Drugs On Purpose

Match.com for Drugs

The Molecules Don't Care

Instead of Finding Them by Accident

The Italian Man's Email

Neither Side Effect Sounds as Bad as Lung Cancer

From Public Data to Acquisition

Turn the Crank

The Legacy Realized (After)

Inflammatix: Drug Repurposing from Public Data

The Cell Paper (After)

Sources

V. ON GARAGE BIOTECH

All the Tools Are Waiting

The Next Amgen Starts in a Garage

Add to Shopping Cart

The Rat Colonoscopy

Wisdom of the Crowd

Companies Are Not the Dark Side

Slide of Significance

Sources

VI. ON MEDICINE

The Physician's Complaint

Imprecise Medicine

Those Silly Fools

Visitation from God

Not Otherwise Specified

Medicine as a Game

The Triple Question

Diabetes Donuts

1,500 Ways to Treat Diabetes

Medicine as Computation

Doctors With Computers

Study Every Damn Thing We Do

What This Medicine Changes

The Memorial Line

Sources

VII. ON AI AND SCALABLE PRIVILEGE

The Doors AI Opens

AI as Scalable Privilege

$2.99 to Explain My Echo

17 Doctors Couldn't Figure It Out