←
Moses Lab10 posts
Moses Lab avatar

Moses Lab

@AMosesLab

Alan Moses Computational Biology Lab at @UofT

◷ Joined August 2017
Sandeep Kambhampati avatar
⟳ Moses Lab retweeted
Sandeep Kambhampati@SandeepKambham2·Jun 8

I'm excited to share our new preprint "Vermeer: Autoregressive generative modeling of microscopy predicts protein localization," a collaboration between @MSFTResearch and @insitubiology! Preprint: biorxiv.org/content/10.64898… Protein localization is fundamental to protein function, but experimental imaging, a key technique to study localization, cannot scale to the entire human proteome across all biological contexts. To address this challenge, we introduce Vermeer, an autoregressive generative model trained on Human Protein Atlas data @ProteinAtlas. Vermeer synthesizes fluorescence microscopy images of proteins conditioned on cell morphology reference stains and protein sequence embeddings (ESM-C). Vermeer can leverage latent information about protein localization in the ESM embeddings to generalize to proteins the model has never seen during training. (1/n)

Andrew Duncan avatar
⟳ Moses Lab retweeted
Andrew Duncan@AgduncanB·Apr 17

Designing enhancers is key for gene therapy but limited by a lack of experimental data. To overcome this, we developed EnhancAR, a generative model trained on 1.7M enhancer homolog families (~230M sequences) to leverage functional information conserved over evolution.

Post attachment
Micaela avatar
⟳ Moses Lab retweeted
Micaela@micaelanonsense·Feb 5

Announcing our preprint: Predicting evolutionary rate as a pretraining task improves genome language model representations: biorxiv.org/content/10.64898… Genome language models (gLMs) have the potential to further understanding of regulatory genomics without requiring labeled data...

Taraneh avatar
⟳ Moses Lab retweeted
Taraneh@taraneh_z·26 Apr 2024

Very happy to share what I've been working on in Barcelona with @BenLehner ! We made many (>400k) mutations in a disordered peptide and the globular domain it interacts with to gain a comprehensive map of how specificity is encoded in a partially fuzzy protein interaction (1/n)

Excited to share @taraneh_z's preprint: A complete map of specificity encoding for a partially fuzzy protein interaction biorxiv.org/content/10.1101/…

Peter Koo avatar
⟳ Moses Lab retweeted
Peter Koo@pkoo562·4 Mar 2024

Do current genomic language models (pre-trained on whole genomes) learn a foundational understanding of biology in the non-coding region of human genomes? A new evaluation led by @AmberZqt suggests not yet! 1/N paper: biorxiv.org/content/10.1101/…

Dmitry Penzar avatar
⟳ Moses Lab retweeted
Dmitry Penzar@dmitrypenzar·21 Jan 2024

Really like the idea from this paper -- biorxiv.org/content/10.1101/… Also, I like the accuracy of the study, e.g authors compare model trained with their procedure (train on augmented data + finetune on real) not only with baseline model (train on real) but on finetuned baseline model

Moses Lab avatar
Moses Lab@AMosesLab·8 Sep 2017

Hi Twitter! This is mostly an automated bot that will publish content directly from our website and blog.