.Jian Zhou is kicking off #MLCB2024 w/ Puffin doi.org/10.1126/scienc…, a unified model explaining the sequence basis of transcription initiation in the human genome. Prediction specificity & the number of insights they deliver for understanding & design are incredible!

Concurrently-developed deep learning models to capture transcription initiation syntax: Anshul Kundaje (anshulkundaje@bluesky) lab's ProCapNet doi.org/10.1101/2024.0… Charles Danko lab's CLIPNET doi.org/10.1101/2024.0… #MLCB2024

#MLCB2024

Big news: I’m soon starting as Professor and Lyda M. Ludeman Endowed Chair in Bioinformatics in Women’s Health Research in the Dept of Biomedical Informatics CU Anschutz Medical Campus. We are recruiting for multiple positions. CU Department of Biomedical Informatics (DBMI) Ludeman Center #WomensHealth news.cuanschutz.edu/dbmi/barbara-s…

Sergey Ovchinnikov Sergey Ovchinnikov is giving an incredibly instructive talk about protein and genomic language models. The illustration of complex ideas slide after slide is just brilliant! [The photos don't do justice] #MLCB2024

📢#MLCB2024 Check out our recent review to break into translational biomedicine! We take a deep dive into methods for comparing & transferring data & knowledge across species, and highlight gaps/challenges. Details on implementation, data, & benchmarks: github.com/krishnanlab/cr…

Three studies showing the difficulty of predicting personalized sequence→expression using genomic deep learning models: Sara Mostafavi Maria Chikina nature.com/articles/s4158… Nilah Ioannidis lab's nature.com/articles/s4158… Katie Pollard lab's doi.org/10.1101/2024.0… #MLCB2024

Our work with Dr. Marylyn D. Ritchie, PhD, Diego Milone and Casey Greene on the Clustermatch Correlation Coefficient (CCC) is finally out in Cell Systems! Article: sciencedirect.com/science/articl… GitHub repo: github.com/greenelab/ccc More details below👇

A refreshing talk by Maria Chikina on building sequence-to-function models that are simpler yet thoughtfully-built incorporating biological phenomena at multiple scales, making these models highly interpretable. See doi.org/10.1093/bioinf… & youtube.com/watch?v=DRRTyP… #MLCB2024

.Jishnu Das is presenting SLIDE for identifying outcome-relevant latent factors in high-dimensional multiomics data. LFs have identifiability guarantees), span scale, & capture linear & nonlinear relationships. nature.com/articles/s4159… #MLCB2024

This was bound to happen: An approach developed in 2004 was just referred to as “old school”! #MLCB2024

Gokul Gowri from Wyss Institute proposed latent mutual information (LMI), a new method that exploits underlying low-dimensional structure in high-dimensional data to approximate MI in high-dim settings w/ realistic sample sizes. arxiv.org/abs/2409.02732 #MLCB2024

.Ava Amini gave an expansive talk on generative protein design & introduced EvoDiff, a general-purpose diffusion framework for programmable, sequence-first protein design. 📜 doi.org/10.1101/2023.0… 💻 github.com/microsoft/evod… #MLCB2024

Another fantastic edition of MLCB comes to close! It was fun to meet wonderful folks in the community & learn cool new techniques. Many thanks to the organizers for a great meeting David A Knowles (@davidaknowles.bsky.social) Sara Mostafavi Anshul Kundaje (anshulkundaje@bluesky) Su-In Lee Gerald Quon James Zou!! #MLCB2024

To understand biological processes, spotting patterns in data is crucial. Traditional correlations can miss nonlinear patterns. Milton Pividori, PhD, and Casey Greene, PhD, introduce a not-only-linear coefficient to help. Learn more ️▶️ bit.ly/47oYj6z Via ScienceDirect

Excited to share our latest preprint! With a great team led by Natalie Sauerwald and @avithemicrobe: “Decomposition of phenotypic heterogeneity in autism reveals distinct and coherent genetic programs.” Our paper sheds light on structures of heterogeneity in autism, and how

Thankful for the Waterman Award for giving me the opportunity to create a unique Community Engagement position for my lab. Apply if you're interested!

.CU Postdoc Assoc is celebrating the @NationalPostdoc Appreciation Week 🎊 #NPAW2024 with a myriad of events through the week! At CU Department of Biomedical Informatics (DBMI), on Sep 19 (12–1p MT), I'm offering: Postdoc ergo proper doc, a workshop on planning & executing an effective postdoctoral experience.