82% of human protein-coding genes now have robust, consistent functional annotations, thanks to the combined efforts of biocurators across the GO consortium, Model Organism Databases, SwissProt/UniProtKB, all based on explicit evolutionary modeling via PANTHER families!
THE PAN-GO #FUNCTIONOME IS LIVE! Read our brand new #openaccess @nature.com paper and explore the PAN-GO Functionome, our most accurate and complete resource ever!
Paper links, press release, how to contribute and more on the @alliancegenome.bsky.social Forums: community.alliancegenome.org/t/8343
Paper links, press release, how to contribute and more on the @alliancegenome.bsky.social Forums: community.alliancegenome.org/t/8343
February 26, 2025 at 10:04 PM
82% of human protein-coding genes now have robust, consistent functional annotations, thanks to the combined efforts of biocurators across the GO consortium, Model Organism Databases, SwissProt/UniProtKB, all based on explicit evolutionary modeling via PANTHER families!
Today's keynote at #ISMBECCB2025 Amos Bairoch talking about 45 years of data curation #swissProt #cellosaurus
July 21, 2025 at 8:02 AM
Today's keynote at #ISMBECCB2025 Amos Bairoch talking about 45 years of data curation #swissProt #cellosaurus
"...breakthroughs like AlphaFold, mRNA vaccine development and antimicrobial resistance tracking would not have been
possible without decades of open, structured data from resources like the PDB, UniProt/SwissProt, ENA and other ELIXIR Core Data Resources"
possible without decades of open, structured data from resources like the PDB, UniProt/SwissProt, ENA and other ELIXIR Core Data Resources"
📣 ELIXIR has published a position paper on the importance of #OpenData infrastructure for 🌍 European competitiveness in life sciences.
Read the news 👉 https://loom.ly/22XW72k
#ResearchInfrastructure #Science4EU
Read the news 👉 https://loom.ly/22XW72k
#ResearchInfrastructure #Science4EU
June 24, 2025 at 9:13 AM
"...breakthroughs like AlphaFold, mRNA vaccine development and antimicrobial resistance tracking would not have been
possible without decades of open, structured data from resources like the PDB, UniProt/SwissProt, ENA and other ELIXIR Core Data Resources"
possible without decades of open, structured data from resources like the PDB, UniProt/SwissProt, ENA and other ELIXIR Core Data Resources"
A #Bioinformatics giant passes away. Very sad to hear this news today. Sad for Switzerland, and for the whole world. Amos Bairoch was a part of the Bioinformatics community for many decades. He will be missed. #Database #Biocuration #SwissProt #Prosite #Cellosaurus |
www.sib.swiss/news/amos-ba...
www.sib.swiss/news/amos-ba...
Amos Bairoch, Swiss pioneer of bioinformatics, passes away
We are deeply saddened to announce the passing of our co-founder and Group Leader Amos Bairoch. Emeritus professor at the University of Geneva, he shaped the development of bioinformatics over more th...
www.sib.swiss
December 2, 2025 at 11:50 AM
A #Bioinformatics giant passes away. Very sad to hear this news today. Sad for Switzerland, and for the whole world. Amos Bairoch was a part of the Bioinformatics community for many decades. He will be missed. #Database #Biocuration #SwissProt #Prosite #Cellosaurus |
www.sib.swiss/news/amos-ba...
www.sib.swiss/news/amos-ba...
Hey #TeamMassSpec,
When you run proteomics on non-human species (mouse, rat, macaque, etc.) — which protein FASTA do you prefer?
Taxonomy-filtered UniProt (all entries)
Reference proteome (SwissProt+TrEMBL)
Ensembl/GENCODE
Something else?
When you run proteomics on non-human species (mouse, rat, macaque, etc.) — which protein FASTA do you prefer?
Taxonomy-filtered UniProt (all entries)
Reference proteome (SwissProt+TrEMBL)
Ensembl/GENCODE
Something else?
September 25, 2025 at 12:41 PM
Hey #TeamMassSpec,
When you run proteomics on non-human species (mouse, rat, macaque, etc.) — which protein FASTA do you prefer?
Taxonomy-filtered UniProt (all entries)
Reference proteome (SwissProt+TrEMBL)
Ensembl/GENCODE
Something else?
When you run proteomics on non-human species (mouse, rat, macaque, etc.) — which protein FASTA do you prefer?
Taxonomy-filtered UniProt (all entries)
Reference proteome (SwissProt+TrEMBL)
Ensembl/GENCODE
Something else?
Concept Bottleneck Language Models For protein design
- Introduce CB-pLM (Concept Bottleneck Protein Language Models) from 24M to 3B, trained on UniRef50 and SwissProt over 718 concepts (including Cluster name, Biological process, and Biopython-derived features, etc.)
arxiv.org/abs/2411.06090
- Introduce CB-pLM (Concept Bottleneck Protein Language Models) from 24M to 3B, trained on UniRef50 and SwissProt over 718 concepts (including Cluster name, Biological process, and Biopython-derived features, etc.)
arxiv.org/abs/2411.06090
December 14, 2024 at 10:29 PM
Concept Bottleneck Language Models For protein design
- Introduce CB-pLM (Concept Bottleneck Protein Language Models) from 24M to 3B, trained on UniRef50 and SwissProt over 718 concepts (including Cluster name, Biological process, and Biopython-derived features, etc.)
arxiv.org/abs/2411.06090
- Introduce CB-pLM (Concept Bottleneck Protein Language Models) from 24M to 3B, trained on UniRef50 and SwissProt over 718 concepts (including Cluster name, Biological process, and Biopython-derived features, etc.)
arxiv.org/abs/2411.06090
Listomics Richard!! Better to have fewer validated targets than a truck load of potential hits. When asked once if we can get the 20,000 protein proteome I said look in SwissProt; they are all there.
November 21, 2024 at 9:38 AM
Listomics Richard!! Better to have fewer validated targets than a truck load of potential hits. When asked once if we can get the 20,000 protein proteome I said look in SwissProt; they are all there.
A lovely tribute to Amos Bairoch from Vivienne Baillie Gerritsen, author of the always interesting Protein Spotlight blog for #SwissProt
www.proteinspotlight.org/back_issues/...
www.proteinspotlight.org/back_issues/...
a tribute
As I crossed Geneva this morning and approached the office, I felt a growing sadness. I realised that, if I was walking in this direction in the first place, it was thanks to Amos Bairoch. Flashback t...
www.proteinspotlight.org
December 18, 2025 at 3:57 PM
A lovely tribute to Amos Bairoch from Vivienne Baillie Gerritsen, author of the always interesting Protein Spotlight blog for #SwissProt
www.proteinspotlight.org/back_issues/...
www.proteinspotlight.org/back_issues/...
5/10 We then ran each structure against AF2-SwissProt via Foldseek, testing for statistically significant enrichment of eukaryotic hits (Fisher's exact test, Bonferroni-corrected).
Result: 908 "isomorphic ESPs" (iESPs) — >3× the previously known ESP count.
Result: 908 "isomorphic ESPs" (iESPs) — >3× the previously known ESP count.
March 5, 2026 at 10:46 AM
5/10 We then ran each structure against AF2-SwissProt via Foldseek, testing for statistically significant enrichment of eukaryotic hits (Fisher's exact test, Bonferroni-corrected).
Result: 908 "isomorphic ESPs" (iESPs) — >3× the previously known ESP count.
Result: 908 "isomorphic ESPs" (iESPs) — >3× the previously known ESP count.
On a separate note: I was surprised to find that none of the non-canonical SwissProt isoforms are included in the official human reference proteome (UP000005640).
Anyone know what’s going on here? 🤔
#proteomics #bioinformatics @pwilmarth.bsky.social il
Anyone know what’s going on here? 🤔
#proteomics #bioinformatics @pwilmarth.bsky.social il
October 1, 2025 at 3:13 PM
On a separate note: I was surprised to find that none of the non-canonical SwissProt isoforms are included in the official human reference proteome (UP000005640).
Anyone know what’s going on here? 🤔
#proteomics #bioinformatics @pwilmarth.bsky.social il
Anyone know what’s going on here? 🤔
#proteomics #bioinformatics @pwilmarth.bsky.social il
1. Planet earth.
2. "Assigned functions": assigned GO terms with experimental evidence codes on Swissprot
2. "Assigned functions": assigned GO terms with experimental evidence codes on Swissprot
October 16, 2025 at 6:58 PM
1. Planet earth.
2. "Assigned functions": assigned GO terms with experimental evidence codes on Swissprot
2. "Assigned functions": assigned GO terms with experimental evidence codes on Swissprot
Isoforms are represented differently in SwissProt vs TrEMBL. Different protein forms have explicit, separate entries in TrEMBL. Those related sequences are rolled into one curated SwissProt record. There's one canonical, representative sequence and others are detailed in the record, not the FASTA.
October 1, 2025 at 3:41 PM
Isoforms are represented differently in SwissProt vs TrEMBL. Different protein forms have explicit, separate entries in TrEMBL. Those related sequences are rolled into one curated SwissProt record. There's one canonical, representative sequence and others are detailed in the record, not the FASTA.
December 2, 2025 at 7:11 PM
The SwissProt subset of AlphaFoldDB (540k proteins) takes even less: under 2 hours (proteins are smaller on average).
But does prediction speed matter? 🤔 It depends. In most cases, probably not. After all, if you have only a few structures, it’s worth waiting for more accurate predictions.
(2/n)
But does prediction speed matter? 🤔 It depends. In most cases, probably not. After all, if you have only a few structures, it’s worth waiting for more accurate predictions.
(2/n)
February 10, 2025 at 1:28 PM
The SwissProt subset of AlphaFoldDB (540k proteins) takes even less: under 2 hours (proteins are smaller on average).
But does prediction speed matter? 🤔 It depends. In most cases, probably not. After all, if you have only a few structures, it’s worth waiting for more accurate predictions.
(2/n)
But does prediction speed matter? 🤔 It depends. In most cases, probably not. After all, if you have only a few structures, it’s worth waiting for more accurate predictions.
(2/n)
There might be a download option to include canonical plus isoforms (however, canonical and isoforms are terms that only apply to the SwissProt entries).
October 2, 2025 at 2:18 PM
There might be a download option to include canonical plus isoforms (however, canonical and isoforms are terms that only apply to the SwissProt entries).
I am deeply saddened learning that my scientific hero Amos Bairoch has recently passed away. He will be remembered as the mastermind behind the SwissProt and many other endeavours in protein bioinformatics. There is a tribute web page at sib.swiss/amos/
Amos Bairoch — In Memoriam · Amos Bairoch remembrances
This page is a place to share memories, stories, and messages in honour of Amos Bairoch.
sib.swiss
January 13, 2026 at 11:36 AM
I am deeply saddened learning that my scientific hero Amos Bairoch has recently passed away. He will be remembered as the mastermind behind the SwissProt and many other endeavours in protein bioinformatics. There is a tribute web page at sib.swiss/amos/
For the longest time SwissProt Human was the go to, maybe five years ago UniProt Human became popular. Now, especially for plasma proteomics, I see folks searching any old crap to boost numbers. It's really quite sad.
September 27, 2025 at 7:56 PM
For the longest time SwissProt Human was the go to, maybe five years ago UniProt Human became popular. Now, especially for plasma proteomics, I see folks searching any old crap to boost numbers. It's really quite sad.
Over the objections of NHGRI, the NIH council decided that NLM would be the exclusive recipient of funding for nucleic acid information & sequences. The protein-sequence-only Uniprot contract that merged PIR & SwissProt was NHGRI's work around to stay in the information resource funding game.
January 3, 2025 at 2:14 PM
Over the objections of NHGRI, the NIH council decided that NLM would be the exclusive recipient of funding for nucleic acid information & sequences. The protein-sequence-only Uniprot contract that merged PIR & SwissProt was NHGRI's work around to stay in the information resource funding game.
Not really - our infrastructure is completely independent of NCBI/NIH.
But some analyses do use copies of NCBI-consolidated data.
If, for example, NIH were to stop updating NR, then European/Japanese/Chinese sister orgs would need to take the lead.
But e.g., SwissProt is made... in Switzerland
But some analyses do use copies of NCBI-consolidated data.
If, for example, NIH were to stop updating NR, then European/Japanese/Chinese sister orgs would need to take the lead.
But e.g., SwissProt is made... in Switzerland
March 2, 2025 at 4:58 PM
Not really - our infrastructure is completely independent of NCBI/NIH.
But some analyses do use copies of NCBI-consolidated data.
If, for example, NIH were to stop updating NR, then European/Japanese/Chinese sister orgs would need to take the lead.
But e.g., SwissProt is made... in Switzerland
But some analyses do use copies of NCBI-consolidated data.
If, for example, NIH were to stop updating NR, then European/Japanese/Chinese sister orgs would need to take the lead.
But e.g., SwissProt is made... in Switzerland
ok! But I mean is the test set (a) just random proteins without codes on Swissprot, or (b) very novel and interesting proteins, or (c) you can't say :)
So does your submission need to run automatically on the test set after deadline? And be all python? Or is docker OK? Sorry, new to kaggle!
So does your submission need to run automatically on the test set after deadline? And be all python? Or is docker OK? Sorry, new to kaggle!
October 16, 2025 at 7:22 PM
ok! But I mean is the test set (a) just random proteins without codes on Swissprot, or (b) very novel and interesting proteins, or (c) you can't say :)
So does your submission need to run automatically on the test set after deadline? And be all python? Or is docker OK? Sorry, new to kaggle!
So does your submission need to run automatically on the test set after deadline? And be all python? Or is docker OK? Sorry, new to kaggle!