Jesper Madsen
@jespergrud
Assistant Professor @ Sect. for Data Science and Statistics @ Dept. of Mathematics and Computer Science @ University of Southern Denmark.
📢 PREPRINT ALERT 📢 Clustering cells is a key step in most scRNA-seq analysis workflows. It's hard, and harder with batch effects. Our take on the problem, JOINTLY, use a hybrid linear and non-linear NMF. The non-linearity gives performance, the linearity gives interpretability
New from us! We found that targeting mitochondrial programming promoted the formation and maturation of stem cell-derived beta cells. Mitos - more than a powerhouse! #T1D biorxiv.org/content/10.110…
Currently at #EISG2024 in Helsinki 🧬 I presented my PhD research, conducted in @susmandrup ‘s lab and co-supervised by @jespergrud at @GenomicsUnit @NATsdu , in a poster session. Last photo by @Hazemibrahimi - who has posted more photos of the conference. Check them out!
Honesty is the best policy ncbi.nlm.nih.gov/pmc/articles/P…
Our watching party getting ready to dive into what will sure be an excellent talk by @MelinaClaussnit 🤯 at the V2F symposium hosted by @novonordiskfond Center @broadinstitute
🧵: How do you design a network that can optimize (edit, transform, ...) the weights of another neural network? Our latest answer to that question: *Universal* Neural Functionals (UNFs) that can process the weights of *any* deep architecture.
The commentary I wrote is now live at Nature Biotech: nature.com/articles/s4158… (free version rdcu.be/dwSSy).
Extremely excited to open the registration for the 2nd Annual Variant to Function Symposium hosted by the @novonordiskfond Center @broadinstitute co-Chaired by @MelinaClaussnit. A stellar (!!) lineup of speakers & virtual+physical attendance. Please share! broadinstitute.swoogo.com/V2F
Very excited to share a new work on AUPRC, AUROC, and class imbalance: arxiv.org/abs/2401.06091 The short of it? Despite widespread belief, AUPRC is not superior to AUROC under general class imbalance and it can exacerbate biases towards subpopulations with higher prevalence.
@afmoeller @jespergrud present JOINTLY, a hybrid linear and non-linear NMF-based algorithm, providing interpretable and robust cell clustering across batches against over-integration. @NatureComms #singlecell #BiotechNatureComms nature.com/articles/s4146…
Exciting news! Our latest research on "Cell type directed design of synthetic enhancers" has been published in @Nature. We've harnessed deep learning to design enhancers with cell-type specificity. Find our paper here: nature.com/articles/s4158…. Highlights below: [1/n]
Is binarization of scATAC-seq data really necessary? The conclusion from our analysis is that a quantitative treatment is in fact beneficial. Now out in Nature Methods! @gagneurlab @fabian_theis nature.com/articles/s4159… Many additions since the preprint 👇(1/n)
EISG Helsinki 2024 | University of Helsinki. Next instalment! Outstanding speakers and an exciting venue. helsinki.fi/en/conferences…