AI’s uncertain future “If this is the final stage of AI, we’re in trouble.” — Alfonso Valencia on the risks and promise of generative models. Continue Reading

Entrevista de Artur Olesch a @alfonsovalencia.bsky.social en #DigitalHealth sobre el incierto futuro de la #IA. Muy recomendable
aboutdigitalhealth.com/2025/11/06/a...

📄Ver publicación del primer estudio de comorbilidades en PNAS: pnas.org/doi/10.1073/...

🔘Plataforma interactiva de red de conexiones entre enfermedades: disease-perception.bsc.es/rgenexcom/

@beatrizurda.bsky.social @alfonsovalencia.bsky.social

#DíaMundialCáncerMama #19Octubre

⭕ El BSC explora las conexiones entre enfermedades como el #cáncerdemama y busca distinguir qué parte de estas relaciones se explica por la #genética o por 𝗳𝗮𝗰𝘁𝗼𝗿𝗲𝘀 𝗮𝗺𝗯𝗶𝗲𝗻𝘁𝗮𝗹𝗲𝘀 o modificables.

🔘Casi la mitad de estas conexiones tiene un origen que 𝘃𝗮➕𝗮𝗹𝗹á 𝗱𝗲𝗹 𝗔𝗗𝗡 y son potencialmente modificables.

Some diseases show up together. Others rarely appear in the same person

This study looked into whether gene activity (from RNA data) can help explain why

The answer: yes - more than we thought

Relevant as we are testing RNA in LC & ME/CFS patients @amaticahealth

Breakdown:

💻🧬'Los vínculos secretos que existen entre enfermedades.
El estudio del BSC representa el mayor esfuerzo hasta la fecha para explicar científicamente las asociaciones clínicas entre enfermedades'

🗞En @innovaspain.bsky.social
➡ www.bsc.es/4kD

@alfonsovalencia.bsky.social @beatrizurda.bsky.social

@growkudos.bsky.social @bsc-cns.bsky.social

Molecular map reveals hidden disease connections Diseases rarely come alone. Many people experience a chain of diagnoses across their lives—for example, smoking-related lung cancer, asthma followed by Parkinson’s disease, or depression alongside lup...

Excited to see our PNAS paper highlighted on Kudos, with an accessible take on the findings.

Disease links are not random—they can be predicted from the expression of our genes.
www.growkudos.com/publications...

@pnas.org @alfonsovalencia.bsky.social
📄 doi.org/10.1073/pnas...

Gene expression maps explain why diseases often occur together This study reveals how gene expression patterns uncover molecular pathways linking comorbidities, enhancing treatment strategies for overlapping diseases.

🧬 New research in PNAS shows how gene expression patterns reveal why some diseases occur together while others don’t.

Scientists uncovered hidden links — with immune pathways playing a major role.

#GeneExpression #Comorbidity #PrecisionHealth

🔗 www.news-medical.net/news/2025090...

Thank you, Luís!!

Patient stratification reveals the molecular basis of disease co-occurrences | PNAS Epidemiological evidence shows that some diseases tend to co-occur; more exactly, certain groups of patients with a given disease are at a higher r...

Great #networkmedicine work by @alfonsovalencia.bsky.social's team on deriving comorbidity networks from RNA-seq data to study complex disease relationships. Thy show that molecular mechanisms are behind many of the known comorbidities (often via immune response).
doi.org/10.1073/pnas...

No le habéis hecho mucho caso a esto, pero es muy, muy bonito. Y parece revolucionario. Seguro que volveremos a oír hablar de esto.

Mil gracias 🤍🧬

This was devastating.

Totally! In our case it wasn’t even about sample size, but a basic textbook statistical concept 🤖. We even increased the sample size to show the results still held under the reviewer’s definition—and still, they wouldn’t budge.

Totally! In our case it wasn’t even about sample size, but a basic textbook statistical concept 🤖. We even increased the sample size to show the results still held under the reviewer’s definition—and still, they wouldn’t budge.

❤️

Muchas gracias!

Muchas gracias por compartir!

El BSC diseña un mapa molecular que descubre vínculos ocultos entre enfermedades Un nuevo método computacional basado en datos de más de 4 000 pacientes y 45 patologías identifica conexiones clínicas conocidas y sugiere asociaciones inéditas con posibles aplicaciones en el diagnóstico y los tratamientos.

Un nuevo método computacional, creado en el @bsc-cns.bsky.social y basado en datos de más de 4 000 pacientes y 45 patologías, identifica conexiones clínicas conocidas y sugiere asociaciones inéditas con posibles aplicaciones en el diagnóstico y los tratamientos.

www.agenciasinc.es/Noticias/El-...

She explains part of the fight over 2 years with an absurd referee (can happen) and an incompetent profesional editor unable to understand even basic statistics - or worse unable to take a decision by him/her self.

But never mind: Beatriz won and the paper is now published in a better journal.

Very happy to get it out.

For scientific & personal reasons this one is special.
Beatriz has done more work and endured the most difficult - and absurd - publications battles I can remember.
Thanks to PNAS for being "normal" and congratulations to Beatriz.

(Beatriz: next one will be easier!)

12/ This was the road.
👉 Here is the science: bsky.app/profile/beat...
📄 Paper: doi.org/10.1073/pnas.2421060122

11/ Endless thanks to those who supported, listened, laughed, and advised: my colleagues, Davide Cirillo, and especially @alfonsovalencia.bsky.social, for his unwavering support throughout this wild journey. @bsc-cns.bsky.social

10/ And yes– I am officially fluent in rebuttals 🥋 It even helped me win Best Talk at ISMB/ECCB 2025 NetBio– for the science, the presentation, and (yes) the Q&A. #ISMBECCB2025

9/ I wouldn’t wish this road on anyone.
But I’m proud we used the struggle to dig deeper– and that’s where we found some of the most interesting science.

🧬 Novel, underdiagnosed links & mechanisms with therapeutic potential
🧍Works even for rare diseases
🌐 A truly useful resource for the community

8/ Science already takes time, I hope to help make it worth it.

Finally, terrified, we sent it to PNAS @pnas.org.
After one round of review, reports came back: supportive. Positive.
Accepted 🎉 🎉

7/ What I learned:
Publishing can be arbitrary.
Some reviewers make up their minds before seeing the evidence.
One reviewer can wield disproportionate power.
Rebuttals must be painfully clear.
Editors often fail to step in, even when the situation is obvious.
Don’t assume fairness in peer review
👇

6/ The results stood firm.
The reviewer did not.

It was tragic. And honestly, a bit comic.

5/ At one point, I was literally making diagrams with dogs 🐕 and dolphins 🐬 to explain a basic concept every colleague understood instantly.

I even tested the reviewer’s hypothesis, which meant redoing everything with a dataset 3× bigger (yes, manually annotated).