1/ BRAKER4 hatched!
The Earth BioGenome Project is on track to sequence ~1.5M eukaryotic species. Every one needs a structural annotation. No Perl monolith was going to survive that. So we rewrote BRAKER from the ground up. github.com/Gaius-August...
I am recruiting graduate students to my genomics-focused lab in EEB at Illinois Urbana-Champaign. Please spread the word if you know any students with an interest in a computational focus to genomics research! evol.mcmaster.ca/brian/evoldi...
Our EEB department at the University of Illinois at Urbana-Champaign is hiring an assistant professor in evo-devo. See go.illinois.edu/EEBAsstProf for details. Please share!
“Aware of my inability to detect AI use with certainty, I implored the students: please don’t use AI. It’s terrible for you. It’s terrible for me, a dystopian experience of spending weeks giving detailed constructive feedback to a machine.” www.forkingpaths.co/p/the-death-...
Electronic sign in a UIUC campus building advertising free ChatGPT use during finals
Finals start tomorrow here at UIUC. You may have seen the NYMag article outlining how pervasive AI has become in student course work, how it's degrading the learning experience. I wrestled with it all semester in my own course. But OpenAI knows the effects, and they known exactly what they are doing
We will greatly miss our friend and longtime colleague Kevin Drum, as will all who knew him and his tremendous body of work. I had the privilege of editing Kevin for a spell, and I learned a lot from that, too. A top-notch thinker and writer, and a true mensch. www.motherjones.com/media/2025/0...
Just learned political blogger Kevin Drum, who I've read daily since his CalPundit days 20+ yrs ago, has died. Along with Friday cat photos, his most enduring legacy may be bringing attention to the links between childhood lead exposure and early adult crime. I will miss his old-school blogging
Everybody (me included) is getting bogged down on the details of events at NIH et al. Fatal mistake. The big picture is Musk/Trump are trying to shut down all gov-funded Medical/Disease research in the US. Have to understand that to have any hope of stopping it. talkingpointsmemo.com/edblog/act-now
And, my @sciencehomecoming.bsky.social-inspired op-ed is online at LNP/Lancaster Online — in print, it's alongside a piece from two geoscientists with local roots on the front page of the Sunday Perspectives section
lancasteronline.com/opinion/colu...
Molecular Ecology Resources journal cover image of the Klumpy algorithm.
If you want to search and/or visualize how a set of long reads lay down on a reference genome -- whether to check a gene annotation, for assembly error, or another reason, Gio Madrigal's Klumpy can help you do it. Excited to see it on the cover! (and online: catchenlab.life.illinois.edu/klumpy/)
Happy to see our latest work out in G3 @genetics-gsa.bsky.social. This is the fourth notothenioid genome our group has assembled so far -- a highly cold-specialized fish with a large complement of antifreeze proteins and an interesting set of chromosomal fusions/rearrangements.
We know that immune genes often show stronger signatures of selection than the genomic average. But what about immune cell types?
In a new preprint
www.biorxiv.org/content/10.1...
we merge single-cell RNAseq (defining stickleback cell types) and population genomic measures of selection...
Scientists, academics, researchers: We’re excited to share that @altmetric.com is now tracking mentions of your research on Bluesky! 🧪
Wow, Iverson's interview with McNeely gives an in-depth feel for the jazz scene in Champaign-Urbana in the early 1970s -- super interesting. As a jazz enthusiast and fan of the early to mid 1970s period, and someone who moved to C-U in 2015, that was fascinating!
Bluesky now has over 20M people!! 🎉
We've been adding over a million users per day for the last few days. To celebrate, here are 20 fun facts about Bluesky:
Just a quick reminder that PhD applications to work on this funded project are due Dec 1. If you're interested please feel free to reach out to me by e-mail.
This post by Tim Lee on his AI Substack is really good. It describes how 3 trends converged to create the AI boom: the (re)discovery of neural networks, the creation of the 1st massive training dataset, and CUDA, which allowed GPUs to be generalized to train models open.substack.com/pub/understa...
New preprint up! We sequenced hundreds of samples from across one of Earth's oldest living organisms - the Pando aspen clone - to understand how mutations accumulate and spread in long-lived clonal organisms. Our results were…surprising. 1/30
Happy to release Stacks v2.67 today. This release updates process_radtags to make it easier to use SRA data and to filter poly-G (error) runs coming from Nextseq/Novaseq machines + bugfixes. We also added a genotype depth filter to populations #RADseq
catchenlab.life.illinois.edu/stacks/
Please share! My lab will be hiring 2-3 postdocs over the coming months to study the genetic basis of behavioral traits in darters using brain transcriptomics & population genomics. First ad is posted here: tamus.wd1.myworkdayjobs.com/TAMU_Externa... #behavior #popgen #evolution #neuroethology
The populations VCF import function is designed only for variant sites. You could re-export the VCF from populations with the alt population map. If you want to keep your filtering, you can use the loci that were not filtered by vcftools to make a whitelist to feed into populations at the same time
If you are interested in the recent attempts to regulate Apple in Europe with the DMA, this is a really interesting (and long) piece by Steven Sinofsky, a major software engineer at Microsoft during its antitrust days (and after). hardcoresoftware.learningbyshipping.com/p/215-buildi...
Yes, there is a flag to output all sites in the VCF file. The populations program will also calculate Pi and Dxy, accounting for variant/invariant sites, by default.
This project was led and implemented by Gio Madrigal (first, first author!) with assistance from Bushra Minhas.
The tool has two modes of operation – you can give it one or more queries, and it will k-merize and search for them in an assembly and the raw reads, or you can turn it loose to scan a long-read genome (and its raw reads) for inconsistent areas of assembly.
We built Klumpy – a combined k-mer search and visualization tool, written in Python – to help us find and verify antifreeze genes in fishes and we have since used it for a lot more.
On the flip side, you may have a shiny new long-read genome assembly and you may want to make sure some area of the assembly is structured properly and supported by the underlying raw reads.
Many of these elements can be complicated in their architecture – like a gene composed of multiple, repetitive elements, or duplicated in tandem arrays – and automated annotation pipelines may miss it.
If you have a sequence that you love – a gene family, transposable element, regulatory feature – you probably spend a lot of time looking for it in different sets of sequence. If so, you might find our latest preprint and software tool, Klumpy, useful. www.biorxiv.org/content/10.1...