This was not easy to build- it required a bunch of Joe's wizardry along with Playwright. But it's done in a way that is literally one-click-and-submitted! I just submitted my first paper with it. This is going to save many, many hours of pain. 3/

Just clone the repo, and use an LLM of your choice, put the files to be uploaded in a directory, and use an LLM to create a .sub file which will have everything needed for the submission. PaperPush already has templates for 20+ journals. Easy to add more. 2/

Looks like Lior Pachter got better impact metrics on his post than you got on yours.

I finally got around to looking at the Bonsai paper which proposes to approximate single-cell dissimilarity maps with tree metrics, which is a perfectly reasonable thing to do. Seems consistent with what we wrote in our "Specious art" paper.

Functional PCA has a long history dating back to the 1990s. I think the contrastive version we introduce via the Rayleigh quotient is going to be very useful. We showcase the power on a bulk RNA-seq time series, finding relevant and interesting differentially variable genes. 13/

The contrastive spatial analysis is incredibly powerful. Examining a colorectal spatial single-cell RNA-seq sample (Visium HD) as target with a matched normal (non-spatial) as background, we are able to delineate tumor cells, and find interesting and relevant gene programs. 11/

In Part II, we develop ρPCA in two ways: adding kernel weighting (from spatial PCA), and also operating in the space of basis coefficients (from functional PCA). The former is great for spatial analysis, the latter for time series. 9/

ρPCA is super useful in anytime one wants to examine differences in variance (rather than just the mean), while performing dim. reduction. Our preprint gives an example using single-cell RNA-seq: we find differentially variable genes (rather than differentially expressed). 5/

Whereas @jameszou.bsky.social's approach to contrastive PCA has a parameter alpha (that can be difficult / impossible to choose), maximizing the RQ (we call this ρPCA) is straightforward, and produced much better results. We show this with many examples. 4/

The Rayleigh quotient for two square matrices A and B is defined as R(v) = (vᵀAv) / (vᵀBv). Maximization of this RQ, which is easy to solve via the (generalized) eigenvalue problem Av = 𝜆Bv. This is illustrated below (a) target and bg, (b) contrastive PCA (c) variance. 3/

For example, consider nf-core at github.com/nf-core/rnaseq. It is distributed under the MIT license. Period. If the authors think there should be more restrictions (e.g. output of a port should match exactly) then they should modify the license the software is distributed under to require that.

CellSweep was motivated by analysis of the 8^3 dataset from the Mortazavi lab, where they noticed contamination in Parse single-cell RNA-seq by assessing marker genes for tissues from multiplexed plates. CellSweep can sweep away these problems as well. 4/

We validate CellSweep in many ways. Here is the comparison to the CellBender validation, although CellSweep is much faster. 3/

CellSweep is also very useful for spatial assays, where we find it can greatly reduce contamination. Interestingly, the lowest quality cells inferred by CellSweep in this VisiumHD dataset are on the border of the image. 2/

Ambient RNA & barcode swapping is a serious issue in single-cell genomics. Tools such as CellBender, scAR, DecontX & SoupX. We have developed CellSweep which is faster (in some cases by a lot) and much more accurate. Extensively tested and benchmarked. www.biorxiv.org/content/10.6... 1/

I used Claude Opus 4.5/4.6 (and a bit of Codex GPT-5.3) to port edgeR to Python. See edgePython github.com/pachterlab/e... This allowed me to develop a single-cell DE method that extends NEBULA with edgeR Empirical Bayes. All in one week. Details in doi.org/10.64898/202...

We're using the term "paracite" to describe such an error by a machine. Certainly it's happening now. E.g. in the case of UMAPs and my paper with Tara Chari, see this rubbish: www.oreateai.com/blog/navigat... 3/

Yet this paper that cites us makes it sound like we approve of UMAP for certain tasks (just not others). That's not true. academic.oup.com/bioinformati... This is obviously bad practice, but it can be hard to know if it's accidental error, sloppiness, or deliberate misrepresentation. 2/

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).

Time for MM... Multi-Modal (single-cell genomics)... flip it around... Wicked Witch! 😈

My student Catherine Felce will be defending her thesis @caltech.edu next week. If you're in the area consider attending; it will be a treat! Cat's recent work: journals.aps.org/pre/abstract... www.biorxiv.org/content/10.1... www.biorxiv.org/content/10.1...

he underlying optimization problems are NP-hard, but for many visualization tasks can probably be solved optimally. It was fun returning to Bryant and Huson's NeighborNet as a good heuristic, but there is certainly room for improvement.

The advantage of organizing with wompwomp is evident in the clarity one sees when tracing, say, the Lannisters vs. Westeros

The improvements produced by wompwomp are evident in the Game of Thrones dataset displayed at the opening of this thread. To the left is the alluvial plot without wompwomp, to the right with wompwomp.

wompwomp sorts columns and blocks within columns to best reveal structure in data. For instance, this is a comparison of clustering algorithms viewed in a randomly organized alluvial plot, fixed columns (only blocks sorted), and fully optimized with wompwomp.

In a new work with Joseph Rich and Conrad Oakes we tackle the problem of how to best organize alluvial plots. We formalize two optimization problems and develop a solution for them based on the neighbornet algorithm, implemented in the program wompwomp: github.com/pachterlab/w...

Glacial area reduced by 41.6% from 1896 to 1921. Between 2015 and 2021, the rate at which Mount Rainier glaciers were losing area was more than two times faster than the rate estimated for the period of 2009 to 2015. irma.nps.gov/DataStore/Do... I took this photo on July 26, 2025.

The confiscation of cell phones from children at schools & summer camps is largely the result of a national campaign by @jonathanhaidt.bsky.social But cell phones are safety devices used for notification & early warnings of active shooters, earthquakes, floods, tornados, tsunamis, wildfires, ..

Congratulations to @delaneyksull.bsky.social on successfully defending his PhD (photo below with his committee: Barbara Wold, @mitchguttman.bsky.social, and another of my former students @hjp.bsky.social). His thesis is on software, tools, and methods development for single-cell transcriptomics.

Congratulations to Anne Kil on successful defense of her PhD! Her thesis focused on engineering solutions for saliva as a biomedical diagnostic, from improved collection asmedigitalcollection.asme.org/medicaldevic... to assessing the potential of saliva multiomics www.biorxiv.org/content/10.1...