Peter

CellRank: consistent and data view agnostic fate mapping for single-cell genomics Nature Protocols, Published online: 29 January 2026; doi:10.1038/s41596-025-01314-wCellRank enables cell fate mapping from single-cell RNA sequencing data in a consistent and scalable manner, relying on different views of the data such as pseudotime or stemness potential, RNA velocity and experimental time points.

New Article! CellRank: consistent and data view agnostic fate mapping for single-cell genomics

👀

Since I started analysing single-cell data (summer 2020; I came from 3 years working on microbiome), I was surprised so little attention was given to feature selection... I've tried several methods since then, tending towards omitting the issue.

I am really looking forward to trying this out!!!!!

Our new paper: Entropy Sorting Feature Selection (ESFS)

A computational framework for gene selection from single cell data that extracts biological signals in noisy data while avoiding artefacts from conventional dimensionality reduction

A thread
www.biorxiv.org/content/10.6...

The time of day for cancer immunotherapy is associated with major outcomes. Early is better. Results from a randomized trial of lung cancer, backs up the importance of our circadian rhythm and immune system
www.nature.com/articles/s41...

Single-cell susceptibility to viral infection is driven by variable cell states

www.sciencedirect.com/science/arti... #MicroSky

"publication systems [should] distinguish between dissemination of results & communication of ideas, and optimize them separately. Results should be in explicit, machine-readable form, while narrative text serves as an interpretive layer for human readers" www.biorxiv.org/content/10.6...

AI hallucinations in science manuscripts are a nuisance. Paranormal citations, or paracites, will be a nightmare.

www.biorxiv.org/content/10.6... (w/ @sina.bio & @lauraluebbert.com).

doi.org/10.48550/arX...

doi.org/10.48550/arX...

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.

AlphaGenome is out in @nature.com today along with model weights! 🧬

📄 Paper: www.nature.com/articles/s41...

💻 Weights: github.com/google-deepm...

Getting here wasn’t a straight path. We discussed the story behind the model, paper & API in the following roundtable: youtu.be/V8lhUqKqzUc

"Evolutionary remodeling of non-canonical ORF translation in mammals"... 1000s of hidden protein-coding regions!

doi.org/10.7554/eLif...

PU.1 inhibition sensitizes stem-monocytic AML to BCL2 blockade https://www.biorxiv.org/content/10.64898/2026.01.20.700677v1

Today I’m happy to release cyto, a tool I’ve developed at @arcinstitute.org to dramatically increase our computational throughput with 10x-flex single-cell processing by more than 16X!

FASTR: Reimagining FASTQ via Compact Image-inspired Representation https://www.biorxiv.org/content/10.64898/2026.01.22.701172v1

Do you know of multilevels models being implemented in any single-cell analysis? Does this apply equally for low n and big n problems?

Temporal tissue dynamics from a spatial snapshot - Nature One-shot tissue dynamics reconstruction can infer changes in tissue composition over time, from single-time-point spatial proteomics of human cancers.

new out in Nature www.nature.com/articles/s41...

Which one is you favourite cNMF implementation/package, Will?

Such a cool analysis. Thanks for sharing!

Antisense transcription can induce expression memory via stable promoter repression - Genome Biology Background The capacity of cells to retain a memory of previous signals enables acquisition of unique fates and adaptation to their environment. The underlying gene expression memory can arise from mu...

⚠️ The final work of two former PhD students Till @tschwammle.bsky.social and Verena @verenamutzel.bsky.social is out!
➡️⬅️ They dissect how memory can arise from antisense transcription using mathematical modelling 💻, genomics 🧬 and synthetic biology ⚒️! link.springer.com/article/10.1...

A tweet from Lior Pachter from July 26, 2024: "I think this paper is a Denial of Peer Review Attack (DOPRA). It's kind of like a DoS (denial of service) attack. There is so much data, so many methods, so much code, so many figures, so many panels, so much supplement, so much text, that it is overwhelming. 18/"

every big data paper with a bold story that is impossible to comprehend, evaluate, and independently verify reminds me of DOPRA. i've come to increasingly appreciate small, unassuming papers with humble conclusions that you can track word for word, data point by data point, assumption by assumption.

Evaluating the practical aspects and performance of commercial single-cell RNA sequencing technologies Abstract. The rapid development of updated and new commercially available single-cell transcriptomics platforms provides users with a range of experimental

academic.oup.com/nargab/artic...

🔬 pertpy: a unified, scalable framework for single-cell perturbation analysis, now out in Nature Methods

Designed for modern perturbation data - CRISPR, drug screens, patient treatments - scaling to millions of cells and 1000s of conditions.
👉 www.nature.com/articles/s41...

SCVI - Estimating null expression levels -- www.nxn.se/p/scvi-estim...

Also, l assume the frequency of finding something meaningful is higher for people that are selective enough in what they try.

The value of theory is undeniable IMO, but how do we preserve it in current Academia?

/end

Of course, I agree resilience is necessary, but too much efforts without deep reasons behind experiments may not be the most intelligent way forward. Unstressed, positive mindsets are more productive (quality over quantity) than exhausted ones.

That being said, IMO extensively iterative approaches drain a team too much. The probability of failure for not-so-thought experiments is higher than for deeply-thought ones.

When it comes to people, raw counts of failures matter.

So, why would someone leading a lab invest their team's efforts in these approaches, rather than iteratively trying stuff until something interesting pops up?

I really think the probability of success (as understood by the system: quantity, paper&grants) is higher for the iterative approach.

Theoretical views that put together multiple experiments and explain results with the zoom out are often seen as if that would be obvious or even known by the community - despite never explicitly described and detailed. It seems to me researchers don't consider theoretical approaches novel/useful.