Florian P Bayer

This is very useful.
Thanks for the effort !!

Proteoforms as the true units of physiological function link.springer.com/ar...

---
#proteomics #prot-paper

Yes. 100% agree.. That is bad MS2 co-isolation and ratio distortion :)

Good question how the *spectra* S/N is calculated by Thermo for the Astral...

Evaluation of a Prototype Orbitrap Astral Zoom Mass Spectrometer for Quantitative Proteomics─Beyond Identification Lists Mass spectrometry instrumentation continues to evolve rapidly, yet quantifying these advances beyond conventional peptide and protein detections remains challenging. Here, we evaluate a modified Orbitrap Astral Zoom mass spectrometer (MS) prototype and compare its performance to the standard Orbitrap Astral MS. Across a range of acquisition methods and sample inputs, the prototype instrument outperformed the standard Orbitrap Astral MS in precursor and protein identifications, ion beam utilization, and quantitative precision. To enable meaningful cross-platform comparisons, we implemented an ion calibration framework that converts signal intensity from arbitrary units to ions per second. This benchmarking strategy showed that the prototype sampled 23.1% more ions per peptide than the original Orbitrap Astral MS. This increase in the ion beam utilization resulted in improved sensitivity and quantitative precision. To make these metrics broadly accessible, we added new metrics to the Skyline document grid to report the number of ions measured in a spectrum at the apex of the elution peak or the sum of ions between the peak integration boundaries. Taken together, our results demonstrate the Orbitrap Astral Zoom prototype as a high-performance platform for data-independent acquisition proteomics and establish a generalizable framework for evaluation of MS performance based on the number of ions detected for each analyte. Data are available on Panorama Public and ProteomeXchange under the identifier PXD064536.

Looking at the recent MacCoss paper, one needs to divide intensities of a OTs by 10x to get to intensities of ITs and Astrals pubs.acs.org/doi/10.1021/...

One would need to multiple with MS1 intensity to get better quantities, because otherwise it depends when the TMT was picked how much intensity (ions/sec) you end up with in that moment of time, how many peptides you sum, etc... TMT is "only" excellent in conserving ratios within a TMT experiment.

I don't see *spectra* S/N ratios reported in table S3.

They have mean vs. STD *protein intensity* over three replicates. I didn't go deep enough to their analysis, but normally people just sum up TMT report intensity to protein level. This intensity is however not indicative of actual quantities.

my understating was that the S/N cutoffs are the sum over all TMT reporters. So a threshold of 1440 S/N for an 18plex would be average 80 S/N per reporter.. are the purple numbers retained spectra?

Thanks for sharing this analysis. Just sad that good signal leads to poor resolution in TOF world.

What I was always wondering is if you could do a normal ion processor/HCD-MS2 scan and then for the MS3 scan do first HCD in the IRM and then select N Fragments (by optics n=1, SPS in the IRM?) and do the second HCD in the processor?

But happy that you guys are thinking about it :) :)

(2/2) For me, the problem with the Astral rather is that it currently does not support MS3. I don't mind running the MS at a lower Hz if I can get much better data from it. Every time I meet someone from thermo, I try to spark the MS3 idea. It hasn't ignited yet ..

The Deu-TMT requires careful planning of which samples go to which channel and requires a Deu/nonDeu-"bridge" sample to remove the Deu-RT effect. Then it works as expected.
(1/2)

A sneak preview is already online in proteomicsDB (www.proteomicsdb.org/analytics/KSR).

If you have questions, comments, or ideas, feel free to reach out.

In summary:
* 133 clinical kinase inhibitors x 5 cell lines
* >17 million peptidoform DRCs
* Critical re-evaluation of existing KSRs
* Proper FDR control for KSRs
* Identification of thousands of high-confidence KSRs
* Exposing how kinases shape cell line models and cancer patients

Chemical proteomics decrypts the kinases that shape the dynamic human phosphoproteome Mass-spectrometry-based phosphoproteomics enables the analysis of thousands of protein phosphorylation events across the human proteome. However, there is a lack of scalable, hypothesis-free, and stat...

Did you ever come across a phosphosite in your proteomics data for which nothing was known? - I bet so!

We have developed a new strategy termed "potency coherence analysis" that leverages the drug potency dimension in decryptM to decode the kinases that shape the human phosphoproteome.

Read more:

The authors used CurveCurator to fit and evaluate all dose–response data across different omics assay types. CurveCurator’s scores let you set custom thresholds and interactively identify high-quality curves in ProteomicsDB.

Mapping drug mechanisms with ProteomicsDB: unified omics and cell sensitivity data at scale Abstract. Proteomic and phenotypic cell sensitivity datasets are increasingly important for understanding chemoproteomics and the underlying drug mechanism

Very happy to see that dose-response curves are now super easy to access for everyone in proteomicsDB. Just a few clicks and you can see at which concentrations your favorite drug engages protein target(s), tinkers with signaling pathways, and inhibits cell growth. Read more: doi.org/10.1093/nar/...

Koina: Democratizing machine learning for proteomics research - Nature Communications Koina is an open-source, online platform that simplifies access to machine learning models in proteomics, enabling easier integration into analysis tools and helping researchers adopt and reuse ML mod...

Exited to share our latest work! Out now in @natcomms.nature.com

Koina aims to transform how #proteomics uses machine learning. You no longer need to be a tech wizard to use ML and now can easily run #ML models. Integrated with FragPipe, Skyline and EncyclopeDIA!

www.nature.com/articles/s41...

A very clever approach to learn and predict MS2 spectra for modified peptides. By augmenting modification encodings and combining them with PROSIT, the new model has essentially generalized to ANY PTM - not just those in the training dataset. Super exciting !!

Great to hear that you like it :)

To your question if that data is already in prDB - decryptM: YES it is ! e.g.
www.proteomicsdb.org/drug/107070/...

also i can recommend the interactive dashboards.html on zenodo to explore the data locally
zenodo.org/records/1609...

Very excited to see this story out in Science Signaling!!

Especially the use of dose-dependent profiling at different time points could clearly separate immediate from late and consequential signaling changes in KRAS-driven (phospho)proteomes.

Yes crazy... I still remember the 1.8 million PDF pages we were printing to manually determine what a real dose-response was in the original decryptM paper. Luckily, this is over now :)

Let me know if you need some tricks for the time-dependent hack and good luck with the paper

At the Kusterlab, we have developed a statistical analysis tool (CurveCurator) to mine these proteome-wide dose-response curves. You may want to give it a try. github.com/kusterlab/cu...
3/3

A 1.5 orders of magnitude dose range is not really wide enough for a full dose-response curve, but if you have such a scaled-up workflow, you could easily extend it to more doses with good dose resolution. 2/3

I really like the regulome-wide profiling in a dose-dependent fashion. Going one step further, I would love to see actual TF dose-response curves in Figure 5. So not interpreting each dose individually as bar, but actually have a curve with potency and efficacy estimates for each protein. 1/3

Assessment of false discovery rate control in tandem mass spectrometry analysis using entrapment - Nature Methods A theoretical foundation for entrapment methods is presented, along with a method that enables more accurate evaluation of false discovery rate (FDR) control in proteomics mass spectrometry analysis p...

Assessing error control is fundamental in mass spectrometry-based proteomics. @bo-wen.bsky.social @maccoss.bsky.social @urikeich.bsky.social et al introduce a theoretical foundation for entrapment along with a method for more accurate evaluation of FDR control.
www.nature.com/articles/s41...

I am looking forward to discuss with with you:

• (phospho)proteome-wide dose-response profiling

• statistical analysis of 180 million curves with CurveCurator

• mapping kinase-resolved activities changes due to all target engagements

• (re-)evaluating the kinase substrate space in humans

The human proteome with direct physical access to DNA Zero-distance photo-crosslinking reveals direct protein-DNA interactions in living cells, enabling quantitative analysis of the DNA-interacting proteome on a timescale of minutes with single-amino-aci...

🚨Our new paper is online🚨
We use zero-distance⚡photo-crosslinking⚡to reveal direct protein-DNA interactions in living cells, enabling quantitative analysis of the DNA-interacting proteome on a timescale of minutes. #DNA #Chromatin #Proteomics
www.cell.com/cell/fulltex...

Unifying the analysis of bottom-up proteomics data with CHIMERYS - Nature Methods CHIMERYS is a spectrum-centric and data acquisition method-agnostic algorithm for the analysis of MS2 spectra. It is capable of deconvoluting any MS2 spectrum, regardless of whether it was acquired by...

One algorithm to rule them all? CHIMERYS bridges the gap — DDA,DIA and PRM — together at last!
With #CHIMERYS, we can now directly compare DDA and DIA data — 🍎 to 🍎 finally made possible.
doi.org/10.1038/s41592-025-02663-w
#KusterLab #WilhelmLab #MSAID #Proteomics

Interesting.. that’s a long time for sure!

We are running the astral mostly without FAIMS and experienced for us normal exploris-like cleaning cycles.

It probably also means that a single publication by one group is not proving anything. Only if multiple independent groups report the same result it becomes trustworthy. Given the reproducibility crisis in biology, maybe this AI fabrication just accelerated a deeper problem, that we need to overcome