Does anyone have a high-res version of NCBI GEO logo (www.ncbi.nlm.nih.gov/geo/img/geo_...)? I hate putting their low-res version in the slides/figures. I've tried to use AI tools for upscaling, but the results aren't perfect (colors are off). Anyone want to take a shot at it? Here is my best take:

An important milestone for me and my lab: fgsea has passed DESeq2 on the list of most downloaded Bioconductor packages. It started as a side project during my PhD, back when DESeq2 was already a staple of RNA-seq analysis. I never even thought it could go this far — but here we are!

In the second case study, we show how CORESH can help interpret a gene signature, originally discovered by the UnPaSt biclustering method for patient stratification, to be associated with smoking—a hard thing to do with conventional gene set enrichment approaches. 7/

However, we also implemented an enrichment analysis procedure to automatically highlight the dataset description terms which are over-represented among the top hits, summarizing the search results. 5/

For example, here are the top-10 hits for the hypoxia gene set (alserglab.wustl.edu/coresh/load/...). Even from the titles, it’s easy to see that many of the datasets are related to hypoxia, as expected. 4/

The key idea is that we can pull out datasets with high levels of gene coregulation, without relying on any experimental metadata. We don’t need to understand what DFO or PGC1a are, or even how the samples are grouped, to see that something happens with hypoxia and OXPHOS here 2/

I'm happy to present our new web-tool CORESH alserglab.wustl.edu/coresh/ for searching public gene expression datasets using gene signature as a query, which was just published in @narjournal.bsky.social doi.org/10.1093/nar/... 1/n 🧬 🖥️

Grant deadlines are approaching, which means I need to stop procrastinating by working on the mascarade package. The new 0.2 version features a better algorithm that should work automatically on most of UMAP/tSNE plots. github.com/alserglab/ma... rpubs.com/asergushiche... 1/4

I like the wording! Reminds of this meme

Want to highlight expression of a marker gene for a certain cell type? Easy. 2/3

And while you argued again about using UMAP plots, I made a fun little package called mascarade, which automatically generate masks for clusters on UMAPs/t-SNE. github.com/alserglab/ma... 1/3

We now also have a companion phantasusLite R package. Among other things, it allows an easy access to RNA-seq counts, compatible with GEOquery package bioconductor.org/packages/rel... 5/n

The major feature of Phantasus is access to around 100K GEO datasets. We support both microarrays and RNA-seqs, thanks to ARCHS4 and DEE2 quantification projects 4/n

For Phantasus we have been able to integrate Morpheus' rich JS-based client with an R backend using OpenCPU server. Now we can explore the data interactively AND do the gene expression analysis with R-based methods: PCA, k-means, limma, DESeq2, fgsea, etc 3/n

Pretty interesting preprint from a neighboring lab. They show a major improvement in complete deconvolution and general NMF problem. I really like the image deconvolution example there, really easy to see qualitative difference in reconstruction accuracy!

Got my genetically engineered fluorescent petunias from light.bio and they are majestic! Highly recommend to anyone in the US. They aren't super bright, but really cool anyway.