Ron Alfa (@Ronalfa) - โปรไฟล์ Twitter

ทวีตที่ปักหมุด

Ron Alfa@Ronalfa·4d

1/ Spatial transcriptomics is among the richest view of human biology that we have: 18,963 genes mapped at subcellular resolution. It's also almost never collected outside of research settings. So we trained a foundation model to generate it from a clinical H&E image alone. Meet TARIO-2. 🧵 noetik.blog/p/tario-2-a-wh…

English

7

31

156

11.1K

Ron Alfa@Ronalfa·1d

100%

Patrick OShaughnessy@patrick_oshag

Alex on why AI drug discovery companies need to generate novel data to succeed: "AI models based on the research that's available is a lot of garbage in and garbage out." "A lot of the recorded literature is actually incorrect. There's been tons of studies that show if you go try to replicate the experiments that are in the literature, you don't even get the same results." "The AI companies that I believe are gonna be most set up for success are the companies with a novel way to generate science tokens that don't exist in the public domain."

QST

2

0

25

2.9K

Ron Alfa@Ronalfa·1d

@parash061 @russianblue2009 @SOUJIJP Achievement unlocked!

English

1

0

1

23

पारश@parash061·1d

@russianblue2009 @SOUJIJP @Ronalfa you work has reached Japanese twitter.

English

1

0

1

22

russianblue@russianblue2009·1d

創薬の世界で、これまでの常識を根底から覆す驚異的なニュースが入ってきました。バイオAI企業のNoetikが、治験で「失敗」した抗がん剤の95%は実は有効であり、AIを活用することで救済できるとの確信を示しています。製薬ビジネスは、1つの薬を世に出すまでに10年以上の歳月と数千億円の費用を投じる、極めてリスクの高い産業です。しかし、多くの有望な薬が最終段階の治験で「期待したほどの平均的な効果が出なかった」として捨てられてきました。Noetikはこの現状を「薬のせいではなく、適切な患者に投与されていなかっただけだ」と鋭く指摘しています。彼らが開発した「TARIO-2」というAIモデルは、画期的です。がん患者が日常的な検査ですでに受けている「標準的な組織切片標本」から、約1万9000もの遺伝子の空間的な広がりを予測し、腫瘍のマップを作り出します。これにより、その薬が本当に効くはずの「隠れた対象者」を、かつてない精度で特定できるようになったのです。この主張を「単なる夢物語」と切り捨てることはできません。世界的な製薬大手であるGSK（グラクソ・スミスクライン）はすでに、肺がんと大腸がんを対象に、5年間で5000万ドル規模のライセンス契約を締結しました。保守的な巨大資本がこれほどの投資を行う事実は、AIによる「敗者復活」がすでに実務のフェーズに入っていることを証明しています。もし、これまで「効果なし」とされた薬たちがAIの目利きによって蘇れば、創薬コストは劇的に下がり、より多くの患者に個別最適化された治療が届くことになります。これまで「治らない」とされていたがんに、かつて見捨てられた薬が特効薬として届く。そんな未来が、AIの力で現実のものになろうとしています。最先端のテクノロジーは、過去の失敗さえも資産に変えてしまう。私たちは今、医療の歴史における大きな転換点に立ち会っているのかもしれません。

日本語

8

143

606

60.5K

Ron Alfa@Ronalfa·1d

@Anonym64382296 Sharing some results at ASCO with Agenus. x.com/Ronalfa/status…

Ron Alfa@Ronalfa

TARIO-2 will make an appearance at #ASCO26 this year. See press release today from our partner @Agenus_Bio. More soon! Artificial intelligence (AI) foundation model as a predictor of efficacy of next-generation checkpoint inhibition with botensilimab (BOT) + balstilimab (BAL) in solid tumors using pretreatment H&E images. businesswire.com/news/home/2026…

English

1

0

1

54

Anonymous746292@Anonym64382296·2d

@Ronalfa Where are your data?

English

1

0

43

Ron Alfa@Ronalfa·2d

Yep, IHC doesn't work. But with our models, H&E is all you need.

Jia (Jenny) Liu MD PhD@JiaJennyLiu

Fascinating ADC clinical trial plenary at #AACR26. Across EGFR and CLDN6 ADCs, target IHC expression alone failed to predict efficacy. We are still selecting patients using a unidimensional biomarker for a multidimensional (🎯⛓️‍💥💣) drug. 🤯 Elegant discussion @rne_md

English

3

2

28

4.4K

Ron Alfa@Ronalfa·1d

TARIO-2 will make an appearance at #ASCO26 this year. See press release today from our partner @Agenus_Bio. More soon! Artificial intelligence (AI) foundation model as a predictor of efficacy of next-generation checkpoint inhibition with botensilimab (BOT) + balstilimab (BAL) in solid tumors using pretreatment H&E images. businesswire.com/news/home/2026…

English

1

2

18

1.5K

Ron Alfa@Ronalfa·1d

@ManelEsteller @ayirpelle @EytanRuppin @AACR Interested in your thoughts on our related release 🙏

Ron Alfa@Ronalfa

1/ Spatial transcriptomics is among the richest view of human biology that we have: 18,963 genes mapped at subcellular resolution. It's also almost never collected outside of research settings. So we trained a foundation model to generate it from a clinical H&E image alone. Meet TARIO-2. 🧵 noetik.blog/p/tario-2-a-wh…

English

0

447

Manel Esteller@ManelEsteller·2d

Wow! Applying Artificial Intelligence #AI in H&E pathology slides to infer spatial transcriptomics data and immune hotness in tumors! @EytanRuppin @AACR #AACR26

English

3

16

100

8.6K

Ron Alfa@Ronalfa·1d

@AlexHarding7 Agree

English

0

414

Alex Harding@AlexHarding7·2d

I am extremely skeptical of the long-term prospects for in vivo gene therapy with viral vectors that do random or semi-random integration.

English

15

8

154

24.5K

Ron Alfa@Ronalfa·2d

If you’re at #AACR26 @MaximeDhainaut @NOETIK_ai talk is going to be wild. What happens when you combine TARIO-2 with scaled IN VIVO CRISPR perturbations + pharmacology?

NOETIK@NOETIK_ai

Headed to #AACR26? Maxime Dhainaut, our Director of Spatial Functional Genomics, will be presenting during the session: "Spatial Biomarkers at Single-Cell Resolution: Mapping Tumor Ecosystems with Proteomics, Transcriptomics, and AI." He'll discuss how Noetik maps human tumor ecosystems on nonclinical tissue to decode complex biology and drive the future of precision medicine. Excited to share our latest insights and connect with the community in San Diego. See you there!

English

0

1

14

2.1K

Ron Alfa@Ronalfa·2d

Super fun doing @latentspacepod with Daniel Bear @recursus. I'm sure there are plenty of hot takes in here, and maybe some accidental previews of upcoming work.

Latent.Space@latentspacepod

🔬 Training Transformers to solve 95% failure rate of Cancer Trials the AI for Science pod is back with @RonAlfa, CEO of @NOETIK_ai, and Daniel Bear, VP Research at Noetik, explaining exactly how their team of top AI x Bio researchers and engineers (shoutout @owl_posting) will use AI to cure cancer, by focusing on key bottlenecks like patient selection, and training large cancer foundation models like TARIO-2, an autoregressive transformer trained on one of the largest sets of tumor spatial transcriptomics datasets in the world... which first required years of blind faith in collecting good data to even get going:

English

0

2

27

4K

Ron Alfa@Ronalfa·2d

@htaneja Much appreciated thoughts on culture in the last section, very good.

English

0

1

466

Hemant Taneja@htaneja·2d

I've been posting less and writing more. I've found it helpful to slow everything down and take the time for proper reflection. Today I'm putting out a review of the last quarter: Growing AI Anxiety Turning LLMs into Computers Anthropic and the Department of War Iranian Conflict The Incoming Techlash It's easy to feel like the response to this is for tech to "get tough" But I want to encourage the next generation to do something simple: keep your soul

English

20

31

250

133.7K

Ron Alfa@Ronalfa·2d

@tetsuotrees Who’s removing some?

English

1

0

1

1.4K

Forrest Cox@tetsuotrees·2d

@Ronalfa In fairness, *SOME* B cells absolutely must go...

English

1

0

7

1.4K

Ron Alfa@Ronalfa·2d

Biotech about to go back to curing cancers instead of 1,000 ways to deplete B cells.

English

12

21

228

24.2K

Ron Alfa@Ronalfa·2d

@AdamKSay Here for it.

English

0

1.3K

Adam Kolom@AdamKSay·2d

@Ronalfa banger

Indonesia

1

0

3

1.4K

Ron Alfa รีทวีตแล้ว

signüll@signulll·3d

in the ai era, if your “roadmap” is > 30 days, you’re simply ngmi.

English

112

93

1.4K

149.8K

Ron Alfa@Ronalfa·2d

@cremieuxrecueil Turns out asking people detailed questions about their symptoms across all organ systems will reveal a good amount of constipation etc.

English

0

2

435

Crémieux@cremieuxrecueil·2d

It's wild knowing that so many drug side effects are basically just people being scared of medicine. Whether it's actually a statin or a placebo, people report almost the same side effects, at similar rates, and they recover from those side effects at similar rates, too.

English

37

91

1.4K

54.7K

Ron Alfa@Ronalfa·2d

@billytcl can discuss a collab! reach out

English

0

69

Billy Lau@billytcl·4d

@Ronalfa Any availability on academic use?

English

1

0

150

Ron Alfa@Ronalfa·4d

1/ Spatial transcriptomics is among the richest view of human biology that we have: 18,963 genes mapped at subcellular resolution. It's also almost never collected outside of research settings. So we trained a foundation model to generate it from a clinical H&E image alone. Meet TARIO-2. 🧵 noetik.blog/p/tario-2-a-wh…

English

7

31

156

11.1K

Ron Alfa@Ronalfa·3d

@AAMortazavi Diagnostic lab*

English

0

1

152