Artificial intelligence in medicine and science

mariovitali said:
At the moment I am using ChatGPT GPT-6, Claude Opus 5 and I am also using POE for other LLMs like Qwen, Google Pro, DeepReasoning etc.

I have not concluded if this will be a business yet, this is a fair question and -hopefully- a fair answer from my part. I do own a patent which I immediately discussed here years ago.
Click to expand...
Thanks for clarifying.

mariovitali said:
If anyone has an objection that witholding the methodology introduces hurdles to medical research I am open -as always- to discuss.
Click to expand...
Yes, I have objections. I view this is an outdated way that will not be the way forward and is less beneficial to patients and open collaboration. If we are happy with things happening behind closed doors then I believe it is probably not the best way forward either, I think in that case you simply try to get someone at the Frontier labs or even just research institutes interested, who have infinite compute and unreleased models that currently make the above methodology look like the stoneage.

But I also completely understand where you are coming from. You have worked on exactly this for many years, much longer before others got involved. I just thing progress has already overhauled these ideas.
 
mariovitali said:
Yes it is an interesting one, I agree. At the moment I have not decided whether I can expose the whole methodology since it can be used to other verticals and thus easily copied but I can provide prompts, reasoning, hypotheses being put forward. I think it's a fair "deal" and it does not pose any hurdles in advancing ME/CFS research. What is more important is the hypotheses being able to be presented and also any critique set forth by AI for any research proposal or hypotheses by other researchers.
Click to expand...
Sounds good to me
mariovitali said:
If anyone has an objection that witholding the methodology introduces hurdles to medical research I am open -as always- to discuss. I do not wish to introduce any hurdles. The biggest hurdle I see at the moment is that these hypotheses cannot be posted and thus cricised or otherwise be evaluated here.
Click to expand...
Understood. I know its probably not what you want to hear but the forum rules are there, I’d just accept this is going to remain the case, move on from focusing too much on that hurdle and look at the other proposals that have been made on how to move forward. There are options available, it’s up to you which you take and what you choose to do with your work.
 
hotblack said:
move on from focusing too much on that hurdle and look at the other proposals that have been made on how to move forward
Click to expand...

Thank you , I am not quite certain on this, but the idea of a subforum is still a possibility or not? @Trish ?

EndME said:
Yes, I have objections. I view this is an outdated way that will not be the way forward and is less beneficial to patients and open collaboration. If we are happy with things happening behind closed doors then I believe it is probably not the best way forward either, I think in that case you simply try to get someone at the Frontier labs or even just research institutes interested, who have infinite compute and unreleased models that currently make the above methodology look like the stoneage.
Click to expand...

To be honest, I do not understand why I just described is a problem. The knowledge is and will be shared, everything else is secondary. I do not wish to derail the thread we can discuss further in PM.
 
Not necessarily a response to any specific comment, but rather a response to the general tenor of this thread:

Personally I am a long way behind the curve in relation to AI, and I endorse the current forum rules, however as things are changing so quickly I do wonder if we need to discuss more how sub groups of members might be able to investigate specific AI based projects outside the main areas of S4ME and then also how they might report back.

People will engage in using AI to investigate ME/CFS related issues, and it seems we have three options:
  • Do nothing, just wait and see
  • Find some way to host specific projects
  • Serve to help members network here in relation to projects happening elsewhere.
[cross posted with @mariovitali ]
 
mariovitali said:
I am not quite certain on this, but the idea of a subforum is still a possibility or not? @Trish ?
Click to expand...
If a member wants to discuss a particular ME/CFS related project with a group of volunteer members in private they can ask the committee to set up a private forum and add those individuals to it. All parts of the forum are subject to forum rules, and private forums are required to have a mod in them to deal with any rule breaches. I haven't asked in the committee, but I would think if all the members of that private forum agree they are there to discuss AI generated ME/CFS related material, that would be OK. The main reason for the rule is to prevent the forum being flooded with misinformation and AI slop.

Another possibility if you only want to discuss a specific aspect with a few people is to do so in a forum private conversation.

I imagine some time in the not too distant future we will need to review the rule. Off the top of my head, one option might be to set up a members only forum that any member can join in specifically for members to discuss hypotheses and ways of exploring them using the more sophisticated AI tools including discussing the output of those tools that I haven't a clue about. From what others with more knowledge than me have said so far, I don't think we've reached that stage yet.
 
A new partnership between CERN and Mistral has just been announced
At the heart of this work is the ambition to develop a single pretrained Large Physics Model – the first foundation model for collider physics.
Click to expand...

It will be trained on billions of simulated proton-proton collisions at CERN’s Large Hadron Collider (LHC), the world’s largest and most powerful particle accelerator, and designed to support multiple tasks, from reconstructing particle collisions and identifying particles to flagging unexpected signals and evaluating future detector designs.
Click to expand...

CERN and Mistral join forces to accelerate AI for science – Home | CERN

home.cern home.cern
 
The accessible human surfaceome
www.biorxiv.org

The accessible human surfaceome

Proteins associated with the extracellular face of the plasma membrane ("surface proteins" or the "surfaceome") are important therapeutic targets because they can direct therapeutics to specific cell types and deliver them into cells. However, experimental and computational databases disagree...
www.biorxiv.org www.biorxiv.org

Abstract (paragraphs added for legibility):

Proteins associated with the extracellular face of the plasma membrane ("surface proteins" or the "surfaceome") are important therapeutic targets because they can direct therapeutics to specific cell types and deliver them into cells. However, experimental and computational databases disagree substantially on which proteins make up the human plasma membrane proteome, and none focus specifically on the set of accessible surface proteins that would be useful for therapeutic applications.

Here, we used an agentic LLM pipeline to curate the accessible human surfaceome from literature. To assess performance, we built SurfaceBench, a manually curated benchmark of 147 proteins on which five reference databases frequently disagree. Our LLM triage agent classified surface status more accurately than existing databases and recovered 1,108 candidate surface proteins not present in any of them. On held-out clinical and viral-entry surface targets, the pipeline recovered 98–100% of established surface antigens.

A literature-anchored deep-dive pipeline run on over 5,000 candidates then returned 1,783 canonical surface proteins and 1,252 with weaker or context-dependent surface evidence. Each queryable record contains 24 LLM-based annotation fields—including confidence, primary localization, and expression breadth—and seven deterministic structural and topological features.

Unlike existing databases, our records trace every claim to the agent's reasoning and to exact quotes from the full-text literature it could access. All records can be browsed in an open-source viewer. These results show that LLMs reasoning over unstructured scientific text can build updatable, bespoke databases that outperform expert-curated resources, although performance remains limited by agent access to full-text literature.
 
More info on the group behind this here https://deliverome.org/
We are a nonprofit startup building an open atlas of the human surfaceome — abundance, specificity, internalization, and routing — to scale targeted delivery beyond the liver.
Click to expand...
They’ve got an API and all data is open and they include all code, build scripts, prompts, the lot.
 
"NIH Websites Are Changing

"Per Executive Order 14434, federal agencies are now using "Super Intelligence" (SI) in place of "Artificial Intelligence" (AI). You may see both terms across NIH content during this transition.

Infographic

Get the Latest Updates and FAQs on the transition
- NIH has launched its new website experience with the
- beta version of NIH.gov and the new thematic hubs.

"This modernized web environment is designed to make it easier to find health information, research, funding opportunities, and other resources.

"NIH's existing websites will remain available during the transition with redirects and notices guiding you to the new environment.

"Existing websites will sunset on a rolling basis as critical and mandated content is verified and the redirects are put in place."
 
More AI powered protein design tools, this time using natural language to describe what functionality you want

Programmable design of functional proteins from natural language, 2026, Dai et al

Dai, Fengyuan; You, Shiyang; Zhu, Yudian; Gao, Yuan; Fu, Lihao; Zhou, Xibin; Su, Jin; Wang, Chentong; Fan, Yuliang; Tan, Qingsong; Ma, Xiaoxiao; Deng, Xianjun; Yu, Letong; Qian, Hui; He, Yan; Ke, Yitao; Zhang, Zhe; Han, Chenchen; Chang, Xing; Zheng, Liangzhen; Wang, Sheng; Ji, Yuan; Hao, Kexin; Cao, Longxing; Wang, Yajie; Zeng, Anping; Wang, Shunzhi; Zhan, Chunjun; Zhao, Jing; Si, Tong; Liu, Jianming; Lu, Hongyuan; Yuan, Fajie

Abstract
Programming biological function—designing bespoke proteins to perform specified molecular tasks—is a foundational goal of molecular engineering. However, current design models remain fundamentally limited: they typically require either natural proteins as starting points for optimization or manual reformulation of functional goals as geometric and sequence-level constraints to guide candidate generation.

Here we introduce Pinal, a 16-billion-parameter foundation model that designs candidate proteins from natural-language descriptions of desired function. Trained on 1.7 billion synthetically annotated protein–text pairs, Pinal links functional intent to protein sequence and structure. In computational evaluations, generated candidates combined high predicted foldability with functional-description alignment and sequence diversity, providing a basis for prioritizing experimentally testable designs.

We applied Pinal to four distinct design tasks—a fluorescent protein, a polyethylene terephthalate hydrolase, an alcohol dehydrogenase and a metabolic H-protein—and experimentally observed the intended function in each case, including catalytic activity for both designed enzymes. Without iterative experimental optimization, a Pinal-designed H-protein increased product titre by 1.7-fold relative to the corresponding native Escherichia coli H-protein in a multi-enzyme CO2 fixation pathway.

These findings support natural language as a high-level interface for candidate generation in protein design, enabling programmable exploration with reduced reliance on manually specified structural or sequence constraints.

Web | DOI | PDF | Nature

See also this discussion https://arstechnica.com/science/202...uage-comes-to-protein-genome-design-software/
 
You must log in or register to reply here.
Back
Top Bottom