Markus Sommer

Check out our just-published AI-based (temporal convolutional network) system, PSAURON, for predicting eukaryotic proteins, from @markusjsommer. We used a similar method in BALROG, our bacterial gene finder. Note: the "P" is silent in PSAURON 🤣 academic.oup.com/nargab/article…

@nickdimonaco @StevenSalzberg1 Since it scores each amino acid in the protein (with a fairly small receptive field), any piece that looks like a protein would likely get a high score regardless of the overall structure, so definitely not the best tool to detect relatively small differences in isoforms

@markusjsommer @StevenSalzberg1 I was reading this thinking it was for prokaryotes and thought this is familiar! Anyway, looks like a very simple (rare thing these days!) approach to a difficult problem. How do you account for 'misassembled/rearranged isoforms' or if you only have one CDS for a gene?

“One Model to rule them all, One Model to find them, One Model to score them all, and in the Paper cite them.” 🧬🧙‍♂️The eye of PSAURON seeks spurious proteins in your genome annotation! New paper with Aleksey Zimin and @StevenSalzberg1 biorxiv.org/content/10.110…

@assaron @StevenSalzberg1 Yep!

@markusjsommer @StevenSalzberg1 P is silent, right?

@StevenSalzberg1 Closed source code? 10 free predictions per day? That seems a bit too evil, even for a dark lord. Unleash the power of PSAURON via “pip install psauron” Code freely available at GitHub github.com/salzberg-lab/P…

Three @biorxivpreprint releases from my lab last week, and one more today! From postdoc @markusjsommer, his new eukaryotic gene recognizer, PSAURON (the "p" is silent), a successor to his bacterial gene finder Balrog #LOTR biorxiv.org/content/10.110…

@JakobHeinz9 @StevenSalzberg1 @NatlAquarium One short year (and one talented former undergrad intern) later we are a bit closer to understanding our delicate and essential coral ecosystem.

@JakobHeinz9 @StevenSalzberg1 Thank you to the @NatlAquarium Animal Care and Rescue Center for raising awareness of coral disease. This collaboration would not have happened without the tour I took in Oct. 2022 (shout-out to Duncan!)

Mystery disease decimating coral reefs off the coast of Florida? Sounds like a job for @JakobHeinz9 and novel metagenomics methods in the @StevenSalzberg1 lab, with guidance from Stephanie Rosales & Lindsay Huebner! Check out the paper! biorxiv.org/content/10.110…

How does alternative splicing regulate protein structure to enable cell-type-specific protein functions? With @momeara, we used AlphaFold2 + scRNA-seq to investigate this question in our latest preprint. biorxiv.org/content/10.110…🧵

Our paper on the human gene annotation catalog CHESS, which now includes all of MANE and a more rigorous definition of genes of all types. With @elapertea @av_sparrow @thesteinegger @markusjsommer @ErdogduBeril @IliaMinkin and others: genomebiology.biomedcentral.com/articles/10.11…

@daniel_c0deb0t Just start working on proteins

absolutely unfair that graphics people always get to have nice renderings to go along with their papers

watch out SpliceAI, there's a new kid on the block, Splam! More accurate at predicting splice sites and much faster too. Check out the code and the pre-print in the thread from @KuanHaoChao below.

ORFanage is new system by @av_sparrow @ErdogduBeril and @elapertea that finds protein-coding regions in RNA-seq assemblies, guided by annotation. Just published today in @NatComputSci: rdcu.be/dibEl

Accurate DTU detection in complex diseases with SPIT by @ErdogduBeril,@stephaniehicks,@StevenSalzberg1,@elapertea! Findings include previously unreported DTU events in 6 genes associated with schizophrenia. Also available on Google Colab -> Learn more at biorxiv.org/content/10.110…

ORFanage is new system by @av_sparrow @ErdogduBeril and @elapertea that finds protein-coding regions in gene annotation, esp. in RNA-seq assemblies. See what it can do in our new preprint biorxiv.org/content/10.110…

I am thrilled to share our Han1 assembly in the @GeneticsGSA journal G3, with @StevenSalzberg1, @elapertea, and @AlekseyZimin5. It’s a great way to start 2023! doi.org/10.1093/g3jour…

first paper of 2023 for us: a gap-free, fully annotated genome of a southern Han Chinese individual, led by @KuanHaoChao, with @AlekseyZimin5 and @elapertea. We used the #T2T CHM13 genome to fill gaps academic.oup.com/g3journal/adva…

The final version of our paper on using protein structure to determine functionality is out in @eLife: doi.org/10.7554/eLife.…. Some of the prettiest figures we've had in any recent paper, like the one here. Led by @markusjsommer, plus @elapertea @thesteinegger et al.

Our paper describing our new, greatly improved human genome annotation is just out in preprint form, by @av_sparrow @elapertea @markusjsommer @ErdogduBeril @IliaMinkin @KuanHaoChao @JakobHeinz9 @alaina_shumate @thesteinegger @park_sukhwan et al. biorxiv.org/content/10.110…