Most proteins in mammalian tissues are cleared in just a few days. That's a remarkable rate of degradation & rebuilding ! Interestingly, degradation rates vary significantly across the cell types within a tissue. www.biorxiv.org/content/10.1... 1/

Recently, the FDA granted Premarket Approval to IsoPSA for prostate cancer diagnosis. This approval reflects a promising direction: ◼️ Improved diagnostic specificity based on protein modifications (proteoforms). ==> 𝐋𝐞𝐭'𝐬 𝐞𝐥𝐞𝐯𝐚𝐭𝐞 𝐨𝐮𝐫 𝐚𝐬𝐬𝐚𝐲𝐬 -- 𝐚𝐧𝐝 𝐭𝐡𝐢𝐧𝐤𝐢𝐧𝐠 -- 𝐭𝐨 𝐭𝐡𝐞 𝐩𝐫𝐨𝐭𝐞𝐨𝐟𝐨𝐫𝐦 𝐥𝐞𝐯𝐞𝐥.

Many things are beyond our control, but we control how we treat other people. We can choose to act with integrity. We can choose to be authentic & empathetic. We can choose to pay attention, help, support and encourage our colleagues. 𝐂𝐡𝐨𝐨𝐬𝐞 𝐭𝐨 𝐛𝐞𝐡𝐚𝐯𝐞 𝐚𝐬 𝐭𝐡𝐞 𝐩𝐞𝐫𝐬𝐨𝐧 𝐲𝐨𝐮 𝐰𝐚𝐧𝐭 𝐭𝐨 𝐛𝐞.

The "publish or perish" mentality is said to lead to low-quality publications. There is truth to that. But I never felt this pressure. For me, publishing for the sake of publishing is a form of perishing — much worse than leaving academia.

A special opportunity for lovers of mass spectrometry proteomics to join a like-minded team at PTI. Join a collaborative initiative to enable direct protein analysis at unprecedented scale, in partnership with leading instrument developers, academics, and industry leaders. 1/2

A recurrent theme in ML: ◼️ Poorly trained models appear to excel at predictions. "Because the sequence and its function are inexorably linked, even a maximally overfit model with no understanding of gene regulation can predict the expression of sequences that are similar to its training data."

🔥 We’re hiring! A rare opportunity to rethink neurodegeneration: 🟦 Develop creative, rigorous approaches to data interpretation toward mechanistic insights. Join our team at PTI to study neurodegenerative disease using direct protein analysis at single-cell resolution. 1/2

The promise of automating boring tasks sounds great! I just used LLMs to help fill in my annual faculty activity report. - The output was not very helpful and would save me very little time. Have you had success with such tasks? Which model would you recommend?

A reminder that proteins are highly dynamic molecules. 🟧 We have made much progress in measuring & predicting static protein structures, but the dynamics that animate life remain challenging to measure & model.

Teaching has been a highlight this semester, intellectually invigorating and grounding. It creates a powerful synergy with research, helping reconnect day-to-day work to the bigger questions and foundational discoveries that shape biomedical research.

Many US corporations have entangled themselves with politics, seeking favor from politicians. This means our purchase choices can powerfully influence political trends, especially if organized into a movement that unites and reflects the popular sentiments of the US public.

Concerning ... and revealing of how superficial referencing practices might be. 1/2

The successes of protein structure prediction are widely known. However, its limitations and failures are less known, especially among non-experts. Most telling are the failures to predict structure types absent from the training data. ⬜️ This is a distinct machine-learning failure, in ... 1/n

Post-translational protein modifications underpin protein functions and dysfunctions. => Naturally, proteoforms enable molecular disease stratification. This resource article maps 145 PTMs and 195 cleavage sites on tau and uses them as biomarkers for tauopathies. 1/2

I’m grateful to be named a COE Distinguished Professor by Northeastern University. This recognition reflects the support, collaboration, and creativity of many students, and colleagues over the years. I’m thankful to be part of a community that values rigorous science, and mentorship.

This review reflects on a journey I’ve been lucky to witness closely: ◼️ The transformation of single-cell proteomics from a fragile idea into a powerful discipline. The technological progress towards deeper coverage and higher throughput is fast and exciting, but what excites me most is ... 1/n

This panel from a recent Nature Biotechnology article is embarrassing. It claims to include representative single-cell proteomics studies, but it misses the largest (and most highly cited) datasets. Simply embarrassing.

Is there a formula for a competitive NIH grant application? This Perspective aims to provide an inclusive roadmap on the elements of NIH funding. Elements of successful NIH grant applications

Discussing research with the innovators beats scrolling a feed. If you’re in single-cell biology, proteomics, mass spec tech, or computational biology, join our next meeting this July. #SCP2026

Single-cell proteomics is moving past descriptive snapshots to quantitative models of cellular regulation and biological mechanisms. Come to #Boston this July to shape the next chapter and have your voice heard. Abstract deadline: Apr 1, 2026. single-cell.net/proteomics/s...

Protein degradation is often considered a final step in gene expression regulation, but it also regulates the first step: Transcription Many TFs, including p53, HIF-1α and c-Myc, are regulated by 𝐝𝐞𝐠𝐫𝐚𝐝𝐚𝐭𝐢𝐨𝐧-𝐠𝐚𝐭𝐞𝐝 𝐜𝐨𝐧𝐭𝐫𝐨𝐥: ⬛️ It has a key advantage. blog.slavovlab.net/2025/08/16/p...

What's larger, a protein or its templating mRNA ? > The mRNA is much larger. ⬛ 𝐀 𝐬𝐞𝐧𝐬𝐞 𝐨𝐟 𝐬𝐜𝐚𝐥𝐞 -- 𝐪𝐮𝐚𝐧𝐭𝐢𝐭𝐚𝐭𝐢𝐯𝐞 𝐢𝐧𝐭𝐮𝐢𝐭𝐢𝐨𝐧 & 𝐜𝐨𝐧𝐬𝐭𝐫𝐚𝐢𝐧𝐭𝐬 -- 𝐢𝐬 𝐚 𝐩𝐨𝐰𝐞𝐫𝐟𝐮𝐥 𝐞𝐧𝐚𝐛𝐥𝐞𝐫 𝐨𝐟 𝐫𝐞𝐬𝐞𝐚𝐫𝐜𝐡. The figure shows myoglobin protein drawn to scale next to its mRNA template. The coding sequence of an mRNA ...

Messenger RNAs with shorter UTRs tend to live longer on average. This may reflect that shorter UTRs statistically contain fewer destabilizing regulatory elements. The relationship depends strongly on cellular context, sequence content, and RNA-binding protein availability.

An elegant approach for quantifying histone PTMs in single cells. It achieves scalability by combining multiplexing and parallelization. I am glad to see creative extensions of nPOP, just as @andrewleduc.bsky.social and I hoped would be enabled by its flexible design. 1/2

This article concludes: “… AI tools seem to automate established fields rather than explore new ones….” Advancing scientific research needs creative approaches and directions. 1/2

𝐖𝐞 𝐨𝐟𝐭𝐞𝐧 𝐭𝐞𝐚𝐜𝐡 𝐭𝐡𝐞 𝐠𝐞𝐧𝐞𝐭𝐢𝐜 𝐜𝐨𝐝𝐞 𝐚𝐬 𝐮𝐧𝐢𝐯𝐞𝐫𝐬𝐚𝐥 𝐚𝐧𝐝 𝐬𝐞𝐭𝐭𝐥𝐞𝐝. 𝐈𝐭 𝐢𝐬𝐧’𝐭. A Nature article reports that even the most basic rules of RNA decoding can be more flexible - and surprising - than often assumed. In the parasite B. nonstop, all 64 codons can encode amino acids in internal regions of mRNA. 1/4

The evolution of NIH funding 1960s: 4 pages / app --> 60 % funding rate 2020s: 100s pages / app --> 10 % funding rate 1/2

Civilization has successfully capitalized on technological benefits while suppressing their downsides. Yet, some technologies can be a net negative: - Leaded gasoline - Chlorofluorocarbons ⬛️ Scientists have special roles in protecting the public & ensuring technologies serve the common good.

Testing whether AI predictions outperform appropriate baselines is essential. Outperforming most baselines is encouraging, but it is not, by itself, a major achievement. Baselines define a starting point, not a meaningful benchmark for excellence.

Many biomarkers and drug targets are cell surface proteins. Is their abundance predictable from mRNA or total protein levels ? Not well. These examples suggest extensive regulation of protein synthesis, degradation and transport. 🔽

Transformative discoveries, such as restriction enzymes, yeast cell-cycle mutants and CRISPR–Cas9, were once considered simply oddities of nature. Let’s adopt more humility about predicting impact, which can truly be known only in retrospect.

The regulation of protein abundance across single cells in a mammalian tissue An excerpt from a presentation that I gave at the MRC Protein Phosphorylation and Ubiquitylation Unit, Dundee University www.youtube.com/watch?v=dnsY...

This view resonates deeply with me: ⬛ Publish fewer claims and more proof Papers need to include fewer claims and more proof to make the scientific literature more reliable What is driving today’s ‘claims inflation’? 1/3

Writing a data processing pipeline can be fast and easy, especially by "vibe coding". Fast and easy is trendy but not the main goal. More important goals include correctness, transparency, robustness and reproducibility. ◼️ Let's prioritize these goals over fast & easy.

The focussed research organizations enabled by @convergentresearch.bsky.social are driving forward an ambitious agenda ! youtu.be/L5W-lV1K_xQ?...

#SCP2026 will be in July. I am looking forward to it !

Biology’s earliest decisions are written in proteins. Using single-cell proteomics, we discovered protein (and functional) symmetry breaking in mammalian embryos long before major transcriptome differences are detectable. cell.com/cell/fulltex...

Signalling by ubiquitination is... ubiquitous ! It regulates virtually all cellular processes in eukaryotes. Bacteria do not encode a functional ubiquitin system, yet some can manipulate ubiquitin signalling in their host. This example 𝐛𝐫𝐞𝐚𝐤𝐬 𝐚𝐥𝐥 𝐫𝐮𝐥𝐞𝐬 ! 1/3

The most common forms of human prion disease are sporadic: Lifetime risk ~1/1,000–1/10,000. Sporadic means the cause is unknown: No known mutations or infections. It's related to proteostasis and aging in ways that are not understood. ⬛ Proteostasis is key to healthy aging.

As I hope for a successful and productive 2026, my thoughts revolve around the core prerequisites: ◼️ They are both simple & challenging: blog.slavovlab.net/2019/12/31/r...

Many articles reported using an antibody ... that didn’t even bind the key protein in Laflamme’s testing. Those articles had been cited over 3,000 times. A key question is: ◼️ 𝐂𝐚𝐧 𝐩𝐫𝐨𝐭𝐞𝐢𝐧 𝐛𝐢𝐧𝐝𝐞𝐫𝐬 𝐚𝐜𝐡𝐢𝐞𝐯𝐞 𝐡𝐢𝐠𝐡 𝐬𝐩𝐞𝐜𝐢𝐟𝐢𝐜𝐢𝐭𝐲 𝐟𝐨𝐫 𝐞𝐯𝐞𝐫𝐲 𝐞𝐩𝐢𝐭𝐨𝐩𝐞 & 𝐩𝐫𝐨𝐭𝐞𝐨𝐟𝐨𝐫𝐦?

We should create a culture that encourages and stimulates creativity, mentoring and supporting the growth of young researchers towards creative intellectuals.

Scientific research can be the most creative human activity. The research of some scientists carries a signature as distinctive as the style of the most original and accomplished artists.

A mission control for frontier technology An ambitious perspective on technological innovation and scaling by @convergentresearch.bsky.social youtu.be/L5W-lV1K_xQ?...

I do not know if the result reflects just collider bias, but collider bias is certainly highly relevant to this analysis. The discussion raises awareness of collider bias, which is a good thing. Collider effects are relevant at the molecular level too ! 1/2

A new study evaluated Alzheimer’s disease neuropathological changes (ADNCs) based on plasma tau phosphorylated at threonine 217. The results suggest a high prevalence of ADNCs: ◼️Prevalence approached 1/4 even in the cognitively unimpaired group aged 70 years or older.

Post-translational modifications on tau isolated from postmortem Alzheimer's disease brains. Although incomplete, these modifications illustrate that the protein products of a single gene are often diverse. Their functional characterization remains an open challenge.

This article suggests that cancer treatment contributes to mutagenesis: "More than 25% of driver mutations in normal tissue exposed to systemic anti-cancer therapy, including in TP53, could be attributed to treatment."

This article is from people who have thought about FROs for years and have experience with what works and what doesn't. I have always appreciated the restraint in defining the niche of FROs in the broader ecosystem; it comes out clearly in this piece. www.essentialtechnology.blog/p/the-future...

The accuracy and precision of protein quantification by MS depend on sampling frequency. An effective way to improve them is ...