Are you using any of our factor models, such as MOFA? 🛵
You might’ve found it challenging to tailor them to your specific use cases - not anymore!
Introducing MOFA-FLEX: a flexible, modular factor analysis framework designed for customizable modeling across diverse multi-omics data scenarios. 1/n
An interesting "what have we been doing all these years?" result from this paper is how sub-optimal the widely-used uniform sampling scheme can be (cluster all @50%, sample from all clusters equally). In contrast, strategies that account for the relative differences in cluster size improve val loss
Single task, lightweight, short-context bp res. profile models often perform on par or outperform current large, multi task, long context models on counterfactual prediction. Much to do to improve.
Bonus: robust, efficient interpretation of syntax
Great collab with @jengreitz.bsky.social lab.
📢 new preprint alert: So so excited to share our analysis on the impact of common and rare variants on single-cell gene expression in blood, using WGS and scRNA-seq data from nearly 2,000 individuals and 5.4m cells as part of TenK10K phase 1 🧬 www.medrxiv.org/content/10.1...
🧵👇 (1/n)
Are you a postgraduate student interested in protein modelling and drug discovery?
We have an exciting opportunity to join our team at GSK for a 6-9 months internship, working on an ambitious cross-department research project. Apply before March 14th!
www.linkedin.com/jobs/view/41...
Very cool paper from Eddie Park and Yi Xing studying the relationship between intron retention QTLs and expression QTLs. Predictably, genetically regulated intron retention can cause changes in gene expression via nonsense-mediated decay (NMD). www.biorxiv.org/content/10.1...
I think the regulatory model (looping etc) first is good with examples that are less extreme and unusual than fto or limb/shh. I think hmgcr and cholesterol is a proximal regulatory effect.
Our ChromBPNet preprint out!
www.biorxiv.org/content/10.1...
Huge congrats to Anusri! This was quite a slog (for both of us) but we r very proud of this one! It is a long read but worth it IMHO. Methods r in the supp. materials. Bluetorial coming soon below 1/
What do GWAS and rare variant burden tests discover, and why?
Do these studies find the most IMPORTANT genes? If not, how DO they rank genes?
Here we present a surprising result: these studies actually test for SPECIFICITY! A 🧵on what this means... (🧪🧬)
www.biorxiv.org/content/10.1...
We are the Stegle Lab: A bioinformatics group advancing computational methods to study molecular variations and their impact on phenotypes. We are jointly hosted at the German Cancer Research Center (@dkfz.bsky.social) and the European Molecular Biology Laboratory (@embl.org) in Heidelberg, Germany.
New work by Andy Dahl and Michal Sadowski on using GxE to study genetics of drug response now out in Cell Genomics www.cell.com/cell-genomic...
Paired CRISPR screens to map gene regulation in cis and trans www.biorxiv.org/content/10.1101/2024.11....
Cross-tissue isoform switches by human transcription factors suggest a widespread regulatory mechanism www.biorxiv.org/content/10.1101/2024.11....
To be clear, we do have plans for scaling, we just kinda expected more than a couple days notice before getting blasted with a million new users a day.
The team is rapidly deploying fixes and new software to adapt. More servers in the mail.
My group's work dissecting the contribution of common variants to rare neurodevelopmental conditions is now out at nature.com/articles/s41..., led by co-first authors Qinqin Huang (not yet on blue sky) and @emiliewigdor.bsky.social . See below for Emilie's tweetorial.
Not enough bioinformatics in your Bluesky feed? We’ve got you covered. Follow us for our latest news, exciting life science research, updates from our data resources, new tools and training resources.
Haven't heard of EMBL-EBI? Take a look at what we’re working on. www.ebi.ac.uk/about/our-im...
Hey, a question for the genetics community. Does genetic fine-mapping work well? How often does it miss?
We usually find that most fine-mapped variants do not fall within coding or regulatory regions. Is it a limitation of epigenomics or a limitation of fine-mapping? Please share your thoughts!
Schematic overview of the proposed mode of action of the newly discovered REX element. Top: An enhancer can activate a gene at a short distance, but not at at long range. Middle: Presence of (C/T)AATTA motifs within an enhancer enable it to act over long distances. Bottom: Coupling a short-range enhancer to the REX element containing the same motifs turns it into a long-range enhancer.
REX - a mammalian "range extender" element that can turn short-distance enhancers into long-distance enhancers.
New preprint from a collaboration led by Grace Bower and Evgeny Kvon.
doi.org/10.1101/2024...
Yes, and also how it informs selection of training datasets. Selecting English language text for English language comprehension, augmenting image datasets by including rotated versions of the same images for image recognition.
I think it already has some physics (eg covalent bond lengths), just less than you might expect (no force field concept). Similar, llms have some knowledge of language (eg the concept of words), just less than you might expect (no verb/noun concept).
Josh Weinstock @joshweinstock.bsky.social at ASHG on inferring GRN from perturbation data
www.biorxiv.org/content/10.1...
Sad to be missing #ASHG23, but check out the talk by the brilliant Wei Zhou talk on Saturday on our new scalable & efficient method for single-cell eQTL mapping!
This is tremendous news, and the quote that "UK Biobank is the world’s most significant resource for health research" is not an exaggeration. Very happy to see it continue to be sustained, especially as we're doing a workshop tomorrow at #ASHG23 on how to use these data!
www.gov.uk/government/n...
Unfortunately I have to miss #ASHG23 this year, but if you are interested in our group's work, do check out these two posters from Ralf Tambets and Krista Freimann:
Presentations from our group at ASHG next week:
Abdullah @abdnahid.bsky.social first #Bioconductor package {easylift} is released in Bioc 3.18
Facilitates genomic liftover using existing Bioc tools:
ranges |> easylift("hg38")
Abdullah is one of the group that signed up for #tidyomics open challenges: github.com/tidyomics
Very happy to see that our review on gene regulatory networks is being featured by Nature Reviews Genetics and that it inspired its cover! 🥳
doi.org/10.1038/s415...