slavov-n.bsky.social - Profile | ThreadSky | a Reddit-style client for Bluesky

We aim to create a meeting venue for interactive discussions among scientists who rarely attend the same meetings. It stimulates creativity & new research directions.

submitted 2 days ago • 0 comments

The program for the 8th single-cell proteomics conference #SCP2025 is up ! Starting in a couple of weeks: single-cell.net/proteomics/s...

submitted 3 days ago • 1 comment

I have read books on the rise of tyrannical demagogues, but I never understood the motivation of their obsequious followers, the people implementing unjustifiable whims and offering silly justifications. What motivates these people ?

submitted 4 days ago • 1 comment

Yesterday, members of @slavovlab.bsky.social were bursting with excitement for the progress they have seen since starting their PhD. Sensitivity is up ~ 1,000-fold. Throughput is up ~ 10-fold. So, I highlighted drivers of this progress: blog.slavovlab.net/2025/05/10/w... 1/2

submitted 5 days ago • 1 comment

Gene regulatory interactions are assumed to be finely tuned. Yet, making (strong) promoters from random DNA sequences is remarkably easy ! De novo evolved promoters in E. coli exhibit low noise by default. 1/2

submitted 5 days ago • 1 comment

Phosphorylation depends both on the sequence and on spatial proximity in vivo between kinases and their substrates. Both of these factors are accounted for in this preprint 1/2

submitted 5 days ago • 1 comment

Single-cell proteomics is gaining momentum. It can analyze the proteomes of ~ 1,000 single cells / day, and we need to ensure high quality cell isolation and sample preparation: ⬛️ These approaches can help: nature.com/articles/s41...

submitted 6 days ago • 0 comments

We want both scale and quality. In just a few days, we can analyze thousands of cells from complex tissues. It's important to select them well and thus minimize artifacts, such as protein leakage. www.nature.com/articles/s41...

submitted 6 days ago • 0 comments

An alternative to tSNE & UMAP for more accurate data visualization: Tree representations for distortion-free visualization and exploratory analysis of single-cell omics data. The trees are constructed to accurately represent true distances between the objects in the high-dimensional space.

submitted 6 days ago • 3 comments

To be more accessible, #SCP2025 will offer a remote broadcast. Attending virtually is free but requires registration: single-cell.net/proteomics/s...

submitted 6 days ago • 1 comment

One way for AI to exceed human performance.

submitted 7 days ago • 0 comments

This project started with an exciting discovery. As @andrewleduc.bsky.social and team interrogated deeper, we found artifactual contributions. So, we developed methods (scp.slavovlab.net/QuantQC) to reduce such artifacts. nature.com/articles/s41... rdcu.be/ekKTq 🧵

submitted 10 days ago • 1 comment

Our paper on the impact of protein leakage due to loss of membrane integrity for single cell proteomic sample prep is out in @natcomms.nature.com! We analyze thousands of primary single cells from murine tracheal tissue and record membrane permeability (sytox-green)🧵 www.nature.com/articles/s41...

submitted 10 days ago • 1 comment

A deep insight into "prompt engineering" and its connection to rubber ducking.

submitted 10 days ago • 1 comment

I tested the FutureHouse AI with some questions. Below is an example: Have CRISPR perturbations of TFs allowed confident inference of TF - target regulation? ✅ It returned relevant references ⚠️ The results offered more cheerleading than critical analysis. platform.futurehouse.org/trajectories...

submitted 11 days ago • 3 comments

AI excels at summarizing encyclopedic knowledge. For original articles, AI tends to inhibit my thoughts. It is smooth and plausibly sounding, but also shallow & mediocre. Generally, it raises the energy barrier for me to write something original that I consider good.

submitted 11 days ago • 1 comment

These papers triumphed over their rejections blog.slavovlab.net/2014/08/15/p...

submitted 11 days ago • 2 comments

Sharp questions & rigorous methods matter.

submitted 11 days ago • 0 comments

The temptation to share popular themes contributes to superficiality. Scientists are affected by this malady too. How do we promote thoughtful, critical and constructive discussions over superficial popularity.

submitted 12 days ago • 0 comments

GPT-like models trained with bio data (mostly RNA) are likely to learn correlation patterns and make useful interpolations (predictions). Since their correlative patterns reflect indirect correlations, extrapolations should be interpreted with caution.

submitted 12 days ago • 1 comment

As a PhD student, I accepted -- without much thought -- that mRNA abundance is called 'gene expression'. I used this sloppy terminology in my research articles, e.g. molbiolcell.org/doi/full/10.... Now it feels strange. Why use a vague phrase when a precise one is available ?

submitted 12 days ago • 0 comments

One of my earliest memories is of my mother telling me to never criticize our president. I remember being shocked. It made no sense to me. Her explanation also did not make much sense. So, I remember it as a strange thing that I could not understand for years.

submitted 14 days ago • 0 comments

Confounders are a major bottleneck to understanding biology. Interactions confounded by unobserved regulators (variables) cannot be causality inferred. Gazillion data points from the observed variables cannot compensate for unmeasured confounders. What can help ... 1/2

submitted 15 days ago • 1 comment

The inflection point that brought about modern molecular biology was a focus on simpler systems that allowed rigorous experiments & conclusions. This drove much progress for decades. In think modern biomedical research can learn from this past experience.

submitted 16 days ago • 0 comments

I am looking forward to this meeting !

submitted 16 days ago • 0 comments

Simpler models outperform GPT-style models of scRNA-seq data. Evaluation of the zero-shot performance of Geneformer and scGPT suggests that these models may face reliability challenges and be outperformed by simpler methods.

submitted 16 days ago • 1 comment

A very interesting preprint from @selbachlab.bsky.social . "Comparative analyses revealed widespread translation of alternative transcripts and uncovered proteolytic processing as a major, underappreciated source of proteome complexity." 1/2

submitted 17 days ago • 1 comment

We measured degradation rate for over 10,000 proteins from over 115,000 peptide sequences which allowed us to explore how the stability of alternative splicing products affects their abundance. We found a strong overall relationship and visualize a few compelling examples!

submitted 17 days ago • 0 comments

"76 percent of voters, and 61 percent of Republicans, said a president should not be able to ignore the Supreme Court." I do not understand why these are not 100 %. The constitution is explicit about it. www.nytimes.com/2025/04/25/u...

submitted 17 days ago • 1 comment

Academia is not about publishing papers. It’s about creating knowledge and teaching it to the world. It’s about asking big bold questions and mentioning the next generation of intellectual leaders. The difference makes a huge difference.

submitted 17 days ago • 1 comment