Mark R@markrobinsonca@{genomic.social,bsky.social}
@markrobinsonca
statistical bioinformatics; canadian/swiss; #methodsmatter #rstats; open science advocate; grumpy towards prestige worshippers, excessive admin and bros.
Interested in benchmarking computational methods in computational biology, generally? Save the date, submit an abstract, join us next March in Ascona! Conference website: sites.google.com/view/ascona2025 Please retweet and/or tell your colleagues!
Weekly recap: What I'm reading this week in de-extinction and patent law, ATAC-seq benchmarking, large animal genomes, SV genotyping with long reads, cell type-specific enhancer prediction, AI in biosecurity, blog.stephenturner.us/p/weekly-recap…
*DifferentialRegulation* out in Biostatistics: doi.org/10.1093/biosta… A Bayesian hierarchical method to identify differences in splicing dynamics (i.e. gene regulation) between conditions. It works on bulk and single-cell RNA-seq. Also on Bioconductor: bioconductor.org/packages/Diffe…
DifferentialRegulation pre-print online -> doi.org/10.1101/2023.0… *DifferentialRegulation* studies how splicing dynamics change between experimental conditions. It is based on a Bayesian hierarchical method (multiple samples) with latent states (for multi-mapping reads).
Absolutely. To me, Gold OA is one of the most discouraging and wasteful aspects of practising science, propagating enormous inequality, gatekeeping, and Matthew effects.
The "pay to publish model" has been terrible for sci publishing. And no, it was not initially popularized by commercial publishers. In fact, many publishers, including numerous society ones, would be happy to go back to subs 😅 (with all the problems it also has)
DESpace2: a novel framework to compare spatial expression patterns between conditions (e.g., healthy vs. diseased), from spatial omics data. Aim: identifying genes with distinct spatial structures (DSP) between groups of samples. 1/3
Looking forward to giving my talk on community-driven benchmarking via Open Problems tomorrow at the first Ascona benchmarking meeting. @markrobinsonca really knows how to put on a conference in this beautiful location! Half of my lab came along :).
Quick update. We have a pretty exciting lineup of topics/speakers in and around benchmarking that will be presented in Ascona at the end of the month: sites.google.com/view/ascona202… There are few registration slots left, so if you are interested to join us, get in touch.
Interested in benchmarking computational methods in computational biology, generally? Save the date, submit an abstract, join us next March in Ascona! Conference website: sites.google.com/view/ascona2025 Please retweet and/or tell your colleagues!
Excited to share our work on selecting validation metrics for single-cell and spatial omics! Learn to evaluate embeddings, graphs, clustering, and spatial domains with biological relevance. Discover poem, our R package with new spatially-aware metrics! doi.org/10.1101/2024.1…
This is getting exciting. Over 20 abstracts submitted. The committee will be looking at them next week. Looking forward to these discussions. There is still time to submit (due Monday 18th). Google form link: forms.gle/v5FJ21oLcJQokw…
Interested in benchmarking computational methods in computational biology, generally? Save the date, submit an abstract, join us next March in Ascona! Conference website: sites.google.com/view/ascona2025 Please retweet and/or tell your colleagues!
Come join us for a 1w @ISBSIB spatial omics workshop in beautiful Lausanne 🇨🇭 > t.ly/ilpUr < featuring @LarsBorm @ahmedElkoussy @markrobinsonca @YvanSaeys @RaSaghaleyn @AnnaCSchaar @MarcoVarrone & more - deadline: Nov 1st.
biorxiv.org/content/10.110… Nice benchmark of single cell "foundation models" (scGPT, scFoundation) and GEARS (a GNN model) further hyped as "virtual cell models" against linear baselines on perturbation prediction. Long-story short: they can't beat the linear baselines. 1/
Before folks jump to the next level with even bigger, more complex models, can u plz come together with others who have expertise on how to set up benchmarks & baselines to really test whether u r actually making progress? 3/
New benchmarking study of 8 pipelines and 5 widely used methods for scATAC data analysis..includes embedding, graph and partition-based metrics and useful recommendations by data complexity and downstream tasks. Snakemake workflow: github.com/RoseYuan/sc_ch… genomebiology.biomedcentral.com/articles/10.11…
We are #hiring! Join our Department of Molecular Life Sciences @UZH_en as Assistant Professor #tenuretrack in "BioImaging and Data Analysis"! jobs.uzh.ch/offene-stellen… Apply until 30th June 2024 at mnf.uzh.ch/bioimaging! #BioVisionCenter #imageanalysis #computervision #AI #ML
UZH has decided to withdraw from the Times Higher Education World University Ranking. The ranking is not able to reflect the wide range of activities in teaching and research undertaken by universities. news.uzh.ch/en/articles/ne…
"In some labs, the culture fosters a dynamic where students who produce desired results are favored, regardless of quality of the work or even research integrity" kamounlab.medium.com/whats-a-toxic-…