Pau Badia i Mompel

@paubadiam.bsky.social

Postdoctoral researcher at @anshulkundaje.bsky.social‬ Machine learning #ML, gene regulatory networks #GRN, single cell and spatial #omics. Previously at @saezlab.bsky.social Core developer of https://decoupler.readthedocs.io/ at @scverse.bsky.social

How can we hope to understand organismal development when it is controlled by huge, complex networks of interacting genes? 
One option is to move away from molecular details and focus on learning representations and rules. Check out the new perspective from me and @jamesbriscoe.bsky.social

James Briscoe@jamesbriscoe.bsky.social · 6mo ago

New Perspective form Rory Maizels & me: "Gene regulatory networks: from correlative models to causal explanations" Gene regulatory networks are supposed to give us mechanistic explanations of development, so why are we drowning in 'hairballs' of statistical correlations? rdcu.be/e7zx7

New Perspective form Rory Maizels & me: "Gene regulatory networks: from correlative models to causal explanations" Gene regulatory networks are supposed to give us mechanistic explanations of development, so why are we drowning in 'hairballs' of statistical correlations? rdcu.be/e7zx7

Gene regulatory networks: from correlative models to causal explanations

Nature Reviews Genetics - In this Perspective, Maizels and Briscoe discuss the limitations of current models of gene regulatory networks and outline solutions to harness data abundance without...

rdcu.be

We are thrilled to share our new pre-print: “System-wide extraction of cis-regulatory rules from sequence-to-function models in human neural development”. S2F-deeplearning models can accurately encode enhancers, yet decoding these models into human-interpretable rules remains a major challenge.

"We find that while TAD boundaries pair more frequently than non-boundary regions, these interactions are infrequent and are uncorrelated with transcriptional activity of genes within the TAD. (...) These results suggest that TAD boundary architecture and gene activity are largely uncoupled"

TAD boundary architecture and gene activity are uncoupled

Topologically associating domains (TADs) are prominent features of genome organization. A proposed function of TADs is to contribute to gene regulation by promoting chromatin interactions within a TAD...

biorxiv.org

We wrapped up an inspiring week with a visit to NVIDIA headquarters right after the scverse conference. Huge thanks to the NVIDIA Healthcare team and @severin7.bsky.social for hosting us and for the great discussions on scaling single-cell analysis and accelerating open science ⚡️

Bild

Nature suggests you use their "Manuscript Adviser" bot to get advice before submitting I uploaded the classic Watson & Crick paper about DNA structure, and the Adviser had this to say about one of the greatest paper endings of the century:

Bild

How do we know whether our AI solutions are up to our standards, for instance in clinical recommendations? We try to contribute to the answer by our recent study on *Benchmarking large language models for personalized, biomarker-based health intervention recommendations* rdcu.be/eMVeL

Benchmarking large language models for personalized, biomarker-based health intervention recommendations

npj Digital Medicine - Benchmarking large language models for personalized, biomarker-based health intervention recommendations

rdcu.be

Which mutations rewire function of regulatory DNA? Excited to share SEAM: Systematic Explanation of Attribtuion-based Mechanisms. SEAM is an explainable AI method that dissects cis-regulatory mechanisms learned by seq2fun genomic deep learning models. Led by @EESetiz 1/N 🧵👇

Bild

I adored writing this piece. It brings together several of the things preoccupying me right now, like chromatin organization and gene regulation. There's so much more to be said on that. Also, these marine critters look gorgeous. www.quantamagazine.org/loops-of-dna...

Loops of DNA Equipped Ancient Life To Become Complex | Quanta Magazine

New work shows that physical folding of the genome to control genes located far away may have been an early evolutionary development.

quantamagazine.org

Been digging through GEO for multiome single-cell data, frustrating how much is unusable 🫠 (missing annotations, frag files, raw data). Uploading properly takes effort, but it makes your dataset so much more impactful... Props to (most of) the authors who fixed things when I reached out, though 👍