A Ground-truth validation of FDR & false localisation control in proteomics – a must-read from Stefan Tenzer’s lab!
The benchmarks are quite enlightening. We are pleased to see our DIA-NN 2.0 excelling in sensitivity - often by a wide margin - while controlling FDR and false localisation rates.
Template definitions (required/recommended/optional), with clear validation rules for times, units, and intervals. Fully integrated with ontologies & more important community oriented. If you want to be part of this, create an issue; do a PR.
What is coming for #SDRF is huge; we have managed in 20 templates most of the relevant experimental design in #proteomics, including non-MS; #crosslinking, #metaproteomics (soil, water, human gut), DIA/DDA; clinical metadata, oncology proteomics; and single cell #SCP github.com/bigbio/sdrf-...
Which Ollama model?
Quick question: Do you have Claude (❤️) or ChatGPT-OpenAI (🔃)? I mean, if you pay for them.
We’re collaborating with analysis tool providers and instrument vendors to standardize file formats for every @pride-ebi.bsky.social submission: github.com/PRIDE-Archiv...
Want your voice heard? Share your feedback by opening an issue on the guidelines. github.com/PRIDE-Archiv... or PR.
In affinity proteomics, SomaScan and Olink, we still have a long way to go toward standardization, but even before that, we need well-defined output files from each technology. This was one of my surprises during the project. Output files vary massively across tools, platforms, etc.
You can contact pride support team. Here, the format description github.com/PRIDE-Archiv.... Facilities some times provide incomplete formats with columns that are not valid.
Big PR to modify and standardize all channels in Proteomics for all quant methods; working towards more standardization in SDRF and PRIDE at the label of the samples. Feedback more than welcome: github.com/PRIDE-Archiv...
Jaajaj, I have to update the model. But to be honest we get more engage in this way; @ajbrenes.com
SDRF 1.1 is almost here 🎉 Big format & tooling updates ahead. Thanks to the community! We’ll keep refining together. One key question just opened: should we remove prefixes for samples, data files & factor values? Join the discussion 👉 github.com/bigbio/prote... or let me know here.
LFQ Benchmark Dataset - Generation Beta: Assessing Modern Proteomics Instruments and Acquisition Workflows with High-Throughput LC Gradients www.biorxiv.org/cont...
---
#proteomics #prot-preprint
We recently developed pmultiqc.quantms.org that does that for multiple tools. But I think this is the future, guys, welcome to the new era of abundance of tools, libraries, and applications, as code is ♾️X cheaper; we can all produce our own ecosystem of tools/libraries/viz components. Love it
I was going more in the direction that you can really review code, AI-guided 😉. Take own data, put it into the package structure, & tested really quickly with not too much effort. At the same time, you can download the paper data & perform analysis with your own tools without too much effort either
As AI proliferates, expect an explosion of tools, libraries, and papers. 😰
Authors: remember that reviewers will also use AI to reproduce, benchmark, and stress-test your work.
If it’s not open, reproducible, and robust, it won’t survive a cup of tea.
I did Software Engineering in the 20 years ago. I always remember Soft. Eng. teachers always talking about this career is not for coding, is for Soft design, Arch, Orchestration, Requirements gathering. Communication with the client, they always said code writing will be gone at some point. 🤯🚀😎
Just when the week is finishing, I accepted a PR to create a #bioconda package for pridepy github.com/PRIDE-Archiv.... Looks like #pridepy is now automatically integrated into pipelines. Thanks to Jonas Scheid for the PR.
#quantms doubles down on #SDRF, eliminating non-standardized input parameters for greater consistency. github.com/bigbio/quant...
🚀 Exciting news for PRIDE users! The new @pride-ebi.bsky.social web submission is launching in February!
Sorry to hear that
Thanks to all collaborators and contributors, especially to @vadim-demichev.bsky.social #OpenMS team, @compomics.com team, and all the core contributors of #quantms. We now have more than 23 contributors, including @microsoft-copilot.bsky.social and other LLM tools.
Other improvements:
- Support for TMT16 & TMT18 correction matrices
- New .dia file format support + improved docs
- Continued DIA-NN & PTM workflow improvements
- ThermoRawFileParser migrated to bigbio nf-modules
- More robustness for cloud / HPC infrastructure runs
#bigdata #cloud #HPC
#onsite (onsite.quantms.org) – phosphorylation site scoring
- Multiple algorithms: #PhosphoRS, #AScore, #Luciphor
- Advanced localization & confidence assessment using pAla method from @andrewrobertjones.bsky.social group
- Reliable PTM interpretation at scale
#phosphoproteomics #PTMs #proteomics
#quantms-rescoring received a major upgrade, making it more powerful and adaptable across datasets and acquisition strategies.
- MS2 transfer learning across instruments & setups (AlphaPeptDeep)
- Improved model #DeepLC and #MS2PIP handling & fixed rescoring ranges
#machinelearning #proteomics
🚀 #quantms 1.7.0 released (#Caracas)
MS2 transfer learning, #DeepLC, #MS2PIP, #AlphaPeptDeep, advanced rescoring, and onsite phosphorylation scoring are here! 🧵
> github.com/bigbio/quant...
> docs.quantms.org/en/latest/in...
> quantms.org/home
#quantms #proteomics #massspectrometry
2026 is bringing #FragPipe to pmultiqc github.com/bigbio/pmult.... We are starting big this year.
Que Viva Venezuela Libre 🇻🇪.
Happy New Year! All the best to everyone. For us Cubans, the wish every year is simple: health. 🎉
This is so annoying. I started analysing some of the patterns of download of PRIDE data to see if we can help the community, for example, to annotate the most downloaded datasets, or put them into fast-transfer storage. Long story short, Im developing an algorithm to remove bots from our web logs.
