Vanja

(1/13) Excited to share the outcome of the IBIS Challenge! The IBIS challenge united dozens of teams across the world in tackling the problem of modeling transcription factor (TF) binding specificity using a diverse collection of experimental datasets for understudied human TFs.

Our paper on LARGE-scale benchmarking of motif discovery tools is published! nature.com/articles/s42...

It was a long, 7 years long journey, which coordinated efforts of 50+ researchers, proud to be on of them.

More results from Codebook about poorly studied TFs are coming soon.

(15/15) MIXALIME is written in Python, getting the stable version is as easy as 'pip install mixalime', and the source code plus tutorial are freely available at github.com/autosome-ru/...

(14/15) We hope you will find MIXALIME and UDACHA useful in your study of regulatory sequence variants.

(13/15) According to stratified LD score regression, these rSNPs correspond to regulatory regions involved in cell type-specific phenotypes. Most importantly, the collection of significant rSNPs can be fully explored at udacha.autosome.org

(12/15) AS-chromatin variants are predominantly located in promoter and enhancer regions and significantly overlap ADASTRA ASBs and GTEx eQTLs.

(11/15) Finally, we used MIXALIME to analyze 5858 chromatin accessibility datasets from gtrd.biouml.org. In the end, we identified >200 thousand allele-specific chromatin accessibility variants.

(10/15) In most cases, MIXALIME outperforms other popular tools for AS calling, offering a good sensitivity/specificity tradeoff.

(9/15) Using heart CAGE-Seq data from Deviatiiarov et al., we benchmarked MIXALIME by calling AS variants in CAGE-Seq and comparing them to allele-specific transcription factor binding from ADASTRA and eQTLs from GTEx.

(8/15) Copy-number variation and aneuploidy are accounted for by fitting a mixture model assuming that reads originate from haplotypes with different copy numbers.

(7/15) MIXALIME also handles reference mapping bias and aneuploidy, see the underlying math in arxiv.org/abs/2306.08287. To counter the mapping bias, MIXALIME uses separate fits for Alt read counts with the fixed number of Ref reads and vice versa.

(6/15) MIXALIME provides a variety of statistical models to fulfill particular use cases, from a standard binomial model to the beta negative binomial (BetaNB) model that accounts for extra overdispersion.

(5/15) MIXALIME is a novel toolbox that uses MIXture models for ALlelic IMbalance Estimation. In the paper, we describe a general workflow from FASTQ files to allelic read counts and SNV-level allele-specific statistics.

(4/15) Technically, the allele specificity is revealed by counting the number of reads supporting each of the alleles and estimating the statistical significance of the observed allelic imbalance.

(3/15) High-throughput sequencing allows tracking chromatin state, gene expression, protein-DNA interactions, and more. Eventually, all methods yield short reads that can be used to call single-nucleotide variants and assess the allele specificity.

(2/15) Check the original Tweetorial describing the respective preprint x.com/halfacrocodi... or follow the thread below.

(1/15) Yet another sweet bioinformatics "software+database" couple from our team:
Meet MIXALIME, a framework for assessing allelic imbalance, and UDACHA, a database of allele-specific chromatin accessibility, read more at www.nature.com/articles/s41...

Last but not least: this update became possible thanks to the experimental data & motif analysis performed within the Codebook/GRECO-BIT collaboration, ibis.autosome.org/docs/about_us

(4/4) Don't hesitate to grab a fresh release from hocomoco.autosome.org and remember that we also provide a fancy online motif scanner, MoLoTool, in all its interactive JS-powered beauty.

(3/4) Also do not forget that HOCOMOCO also provides motifs for orthologous mouse TFs, >800 of those are covered in v13.

(2/4) v13 covers >1100 of ~1600 human TFs with >1600 primary motifs and subtypes. Since v12 we also provide a reduced non-redundant set of motifs, which are often shared between TFs with similar DBDs.

(1/4) Thrilled to announce another major release of the HOCOMOCO motif collection, well-known for its silly name and rigorous approach to constructing and benchmarking DNA sequence motifs recognized by human and mouse transcription factors.

hocomoco.autosome.org

A generative framework for enhanced cell-type specificity in rationally designed mRNAs mRNA delivery offers new opportunities for disease treatment by directing cells to produce therapeutic proteins. However, designing highly stable mRNAs with programmable cell type-specificity remains ...

The first preprint of 2025! Together with Matvei, @halfacrocodile.bsky.social, & our amazing team, we are excited to share PARADE: an AI framework for designing mRNA UTRs with enhanced cell-type specificity & stability. www.biorxiv.org/content/10.1...

(6/6) 👏 Once again, congratulations to the winners and their mentors including @callitmagic.bsky.social y.social @yaronorenstein.bsky.social @salimovdr.bsky.social @german-roev.bsky.social

(5/6) 🏁 This is the finishing line of the challenge but more remains to be done to analyze the solutions in detail and perform extra tests against other models for the post-challenge manuscript.

(4/6) 🗣️ Check the IBIS website (ibis.autosome.org/challenge_te...) to see the detailed rankings and read the short interviews with the winners!

(3/6) 🏆 We congratulate the runner-up teams: Medici, Salimov & Frolov lab (Russia), callitmagic (Armenia); and, most importantly, the winners: teams Bench Pressers (US), mj (Russia), and Biology Impostor (Israel).

(2/6) 😮 Since the launch, more than 100 teams registered to participate, and 19 teams (with members from 12 countries across the globe) survived till the final stage. The complete IBIS data on Zenodo: zenodo.org/records/1418... and zenodo.org/records/1417...

(1/6) 🐦‍🔥 In IBIS #ibischallenge, we challenged teams from all over the world to decipher the DNA recognition code of human transcription factors. The IBIS Final Conference took place on November 27, 2024. Recordings and slides: disk.yandex.ru/d/82FEnwPn15...

A brief reminder: the IBIS conference is going to happen on Nov 27 = today (or tomorrow, depending on your timezone). Grab the Zoom link at ibis.autosome.org/news and enjoy the show!