Just released: SemiBin v1.2.0 Thanks to Florian Plaza Oñate, we now include a chicken caecum model semibin.readthedocs.io/en/latest/what…

New version of SemiBin released and this includes a big change Version 1.3 introduces a self-supervised mode in addition to the semi-supervised mode that was the basis of SemiBin before

SemiBin v1.3.0 introduced what we think is a big improvement in the results by using self-supervised learning For now, though, we wanted to make self-supervised learning optional as we are still benchmarking it x.com/luispedrocoelh…

2022: A Year Full of Amazing AI papers - A Review 🚀 A curated list of the latest AI breakthroughs by release date with a clear video explanation, article, and code. Finally done! Stay tuned for the video recap next week 😎 github.com/louisfb01/best… #ai #research

Luis Pedro Coelho Finally, I have time and checked the SemiBin result, which improved by ~30% recovered MAGs, incl. both total MAGs (21 - 32%) and HQ-MAGs (27 - 37%). Really impressive! 🍾🎄🎄 Ps. the long-read datasets were from wastewater.

SemiBin2: self-supervised contrastive learning leads to better MAGs for short- and long-read sequencing biorxiv.org/content/10.110…

SemiBin2 preprint online biorxiv.org/content/10.110…

A ~100 pages survey on pretrained foundation models: from BERT to ChatGPT. Has a nice structure for reviewing pretrained models in addition to challenges and open problems. arxiv.org/abs/2302.09419

Here we go! Microsoft introduces a multimodal large language model called Kosmos-1. Achieves great performance on language understanding, OCR-free NLP, perception-language tasks, visual QA, and more.

Top ML Papers of the Week (April 3 - 9): - Segment Anything Model - SegGPT - A Survey of LLMs - Instruction Tuning with GPT-4 - 8 Things to Know about LLMs - Summary of ChatGPT/GPT-4 Research ...

Congratulations to Shaojun Pan as the SemiBin2 manuscript got accepted in the Proceedings track of #ISMBECCB2023 (and will be published in Bioinformatics) See you in Lyon! biorxiv.org/content/10.110…

Strobealign v0.11.0 is out with some big updates github.com/ksahlin/strobe… . Most substantially, strobealign has gotten a new indexing layout. This substantial improvement was contributed by Shaojun Pan (Shaojun Pan) (supervised by Luis Pedro Luis Pedro Coelho). (1/3)

SemiBin2 manuscript is out! tl;dr: SemiBin2 can do MAGs as well or better than SemiBin1 at a fraction of the (computational) cost and also supports long-reads academic.oup.com/bioinformatics…

. Yazhini Arangasamy presented McDevol at the microbiome session in #ISMBECCB2023 and included this very nice historical overview

We just released SemiBin 2.0.0! The previous version was already "SemiBin2 beta," so no major changes, but - we now recommend you use the SemiBin2 command - a few bugs were fixed - we added more internal error checks and better error messages pypi.org/project/SemiBi…

Preliminary results from a side project on the most solved problem in bioinformatics: How do you map short reads to a database?

strobealign v0.13.0 is out. It has built-in coverage estimation*, faster indexing**, and higher accuracy on shortest reads (50-75nt) from parameter optimization. Full log here github.com/ksahlin/strobe… Credits to Shaojun Pan, Luis Pedro Coelho, Adam Lugowski and Marcel Martin (@[email protected])

An impressive advance in metagenomic contig binning by Luis Pedro Coelho and Shaojun Pan enabled by strobealign.

If you have questions about using SemiBin, Small proteins and antimicrobial peptides, NGLess, the Global gene catalog (GMGC), or anything else metagenomics related, you can join our Open Office Hours First session in 30 minutes and the Zoom registration issue is fixed!

Latest preprint from the lab (and the last chapter from Shaojun Pan's PhD thesis) Making strobealign nimbler and faster! biorxiv.org/content/10.110…