Also, I would like to add non-technical points to this discussion. I have seen enterprise software companies offer free solutions for over a decade for scientific labs, and then charging them or just stopping service provision. I strongly believe that a FOSS distro can offer a state of art solution for science, not only concerning infrastructure. (k8s, and so on), but to help and enable actual scientific development on their laptops and lab computers. That is the core of this request and part of the effort to create digital sovereignty.
Priscila. On Tue, Sep 29, 2026 at 2:19 PM Priscila Gutierres <[email protected]> wrote: > Hello Ankur, > > Regarding the Python Packages, I agree with you. Since the bioinformatics > community relies on conda, mamba, and standard virtualenvs for Python > tooling, it is not the SIG's aim to compete with them. > Lemme give a concrete example. Prof. Matej Lexa, at Masaryk University in > Brno teaches bioinformatics and develop tools like TE-greedy-nester > (structure-based detection of LTR retrotransposons) that depend on BLAST+, > LTR_FINDER, and GenomeTools. > None of these dependencies are packaged in Fedora today, forcing his team > to distribute a pre-configured Ubuntu VM and old Linux distributions that > contain those packages. > He previously recommended Bio-Linux as the go-to platform for > computational biology researchers, but Bio-Linux has been discontinued > since 2017. So we have a real gap here. > In HPC and shared computing environments, common in universities, the > users often don't have root access and can't always use conda. So there is > a need to keep well-maintained RPMs for the core compiled toolchain, which > would make a real difference for researchers and students. > I would like to list a initial high-priority list tools, focused on > genetics: > > * BWA, BWA-MEM2, minimap2, HISAT2, Bowtie2, STAR > * samtools, bcftools, htslib, bedtools > * BLAST+, HMMER, DIAMOND > * SPAdes, MEGAHIT > * RepeatMasker, LTR_FINDER, GenomeTools > * IQ-TREE, FastTree, RAxML > * FreeBayes > > Some of these may already be in Fedora. > So, part of the SIG's initial work would be auditing what exists, what's > outdated, and what's missing. > The scope is largely complementary rather than overlapping. The > bioinformatics/genomics toolchain is distinct from computational > neuroscience, > but I'd love to collaborate on shared infrastructure, packaging practices, > and base libraries where our domains intersect. > > I also see the SIG as a coordination point for documentation (e.g., "how > to set up a genomics workstation on Fedora"), a curated container/toolbox > image with the essential compiled tools, and a contact point for > researchers hitting packaging issues. > That's all I have for now. I would really like to hear your thoughts. > > Priscila. > > On Mon, Sep 28, 2026 at 10:27 AM Ankur Sinha <[email protected]> > wrote: > >> On Sun, Sep 27, 2026 17:41:18 +0200, Priscila Gutierres via devel wrote: >> > Hello, >> >> Hello! >> >> > >> > I propose creating a Bioinformatics SIG to improve the state of >> bioinformatics >> > and computational biology tooling in Fedora. >> >> Would you have a list of use-cases and/or packages that we think need to >> be include in Fedora? >> >> We have the NeuroFedora SIG which does package some computational >> biology software, but mostly ones that are related to computational >> neuroscience: >> >> https://neuro.fedoraproject.org >> >> We're currently at ~400 packages (!) >> https://src.fedoraproject.org/group/neuro-sig >> >> Of late, we've been reducing the number of packages we maintain because: >> >> - there's a lot of software, and the few of us can't keep up, so we need >> to (ruthlessly) prioritise >> - for lots of packages, like python packages, we don't have any evidence >> to suggest that researchers prefer system packages over installing >> directly from pypi in virtualenvs---indeed virtualenvs seem to be the >> norm from anecdotal evidence >> >> We had a comp-neuro lab image that we discontinued for similar reasons. >> >> Our focus is now on packages that are non-trivial to build/compile for >> users. For Python packages, unless they are necessary to package, we >> have some regular tests to see if they can be installed off pypi, and we >> note these in our docs and report upstream: >> >> https://ankursinha.fedorapeople.org/neurofedora/package-status/latest/ >> >> >> -- >> Thanks, >> Regards, >> Ankur Sinha "FranciscoD" (He / Him / His) | >> https://fedoraproject.org/wiki/User:Ankursinha >> Time zone: Europe/London >> >
-- _______________________________________________ devel mailing list -- [email protected] To unsubscribe send an email to [email protected] Fedora Code of Conduct: https://docs.fedoraproject.org/en-US/project/code-of-conduct/ List Guidelines: https://fedoraproject.org/wiki/Mailing_list_guidelines List Archives: https://lists.fedoraproject.org/archives/list/[email protected] Do not reply to spam, report it: https://forge.fedoraproject.org/infra/tickets/issues/new
