Davide CIttaro

Do you have some refs to share?

We had a strong enrichment in shorter sequences when testing AVITI, is this something other have noticed?

Problem is that in a combinatorial barcoding experiment we basically sequenced empty artifacts (same library on illumina was legit)

It seems that Cicero is only slightly better than tossing a coin 😨
Also, whatever the approach it seems there’s a huuuuge room for improvement.

3D printing and bioprinting for miniaturized and scalable hanging-drop organoids culture Three-dimensional (3D) cell culture systems rely on the manipulation of a biologically derived matrix, typically soluble Basement Membrane Extract (sBME), in which cells or cellular aggregates, such a...

From the lab next to ours, cool device for HT organoid culture, testing and screening.

SimpleFold: Folding Proteins is Simpler than You Think Protein folding models have achieved groundbreaking results typically via a combination of integrating domain knowledge into the architectural blocks and training pipelines. Nonetheless, given the suc...

arxiv.org/abs/2509.184...

Or that it doesn’t need NVIDIA hardware

I can’t tell if it’s more interesting the approach and results (good predictions+ensembles) or the fact it’s efficient and requires less energy to run. Or both.

scKAN: interpretable single-cell analysis for cell-type-specific gene discovery and drug repurposing via Kolmogorov-Arnold networks - Genome Biology Background Analysis of single-cell RNA sequencing (scRNA-seq) data has revolutionized our understanding of cellular heterogeneity, yet current approaches face challenges in efficiency, interpretability, and connecting molecular insights to therapeutic applications. Despite advances in deep learning methods, identifying cell-type-specific functional gene sets remains difficult. Results In this study, we present scKAN, an interpretable framework for scRNA-seq analysis with two primary goals: accurate cell-type annotation and the discovery of cell-type-specific marker genes and gene sets. The key innovation is using the learnable activation curves of the Kolmogorov-Arnold network to model gene-to-cell relationships. This approach provides a more direct way to visualize and interpret these specific interactions compared to the aggregated weighting schemes typical of attention mechanisms. This architecture achieves superior performance in cell-type annotation, with a 6.63% improvement in macro F1 score over state-of-the-art methods. Additionally, it enables the systematic identification of functionally coherent cell-type-specific gene sets. We demonstrate the framework’s translational potential through a case study on pancreatic ductal adenocarcinoma, where gene signatures identified by scKAN led to a potential drug repurposing candidate, whose binding stability was supported by molecular dynamics simulations. Conclusions Our work establishes scKAN as an efficient and interpretable framework that effectively bridges single-cell analysis with drug discovery. By combining lightweight architecture with the ability to uncover nuanced biological patterns, our approach offers an interpretable method for translating large-scale single-cell data into actionable therapeutic strategies. This approach provides a robust foundation for accelerating the identification of cell-type-specific targets in complex diseases.

I knew it was only a matter of time before KAN made into single cell!

It’s been a true pleasure

Can I suggest a couple?

A subset of transposable elements as mechano-response enhancer elements in controlling human embryonic stem cell fate - Nature Cell Biology Sun et al. identify a subset of transposable elements that serve as mechano-response enhancer elements that control gene expression and human stem cell fate.

What?

A hierarchical, count-based model highlights challenges in scATAC-seq data analysis and points to opportunities to extract finer-resolution information - Genome Biology Background Data from Single-cell Assay for Transposase Accessible Chromatin with Sequencing (scATAC-seq) is highly sparse. While current computational methods feature a range of transformation procedures to extract meaningful information, major challenges remain. Results Here, we discuss the major scATAC-seq data analysis challenges such as sequencing depth normalization and region-specific biases. We present a hierarchical count model that is motivated by the data generating process of scATAC-seq data. Our simulations show that current scATAC-seq data, while clearly containing physical single-cell resolution, are too sparse to infer true informational-level single-cell, single-region of chromatin accessibility states. Conclusions While the broad utility of scATAC-seq at a cell type level is undeniable, describing it as fully resolving chromatin accessibility at single-cell resolution, particularly at individual locus level, may overstate the level of detail currently achievable. We conclude that chromatin accessibility profiling at true single-cell, single-region resolution is challenging with current data sensitivity, but that it may be achieved with promising developments in optimizing the efficiency of scATAC-seq assays.

Among other things, scATAC suffers the inefficient tagmentation process. I can’t agree more, we have some sc data at high coverage and it seems that the number of events per cell is by far lower than expected

🤯

The church of the holy trinity: endoderm, mesoderm and ectoderm!

IDK, but this was a few posts below yours in my feed

bsky.app/profile/adam...

Analog optical computer for AI inference and combinatorial optimization - Nature An analog optical computer that combines analog electronics, three-dimensional optics, and an iterative architecture accelerates artificial intelligence inference and combinatorial optimization in a s...

hey @aoc.bsky.social, they named a new AI computer after you
www.nature.com/articles/s41...

Venice, the Pink Cloud by Paul Signac

A wonderful UMAP!

How do Probabilistic Graphical Models and Graph Neural Networks Look at Network Data? Graphs are a powerful data structure for representing relational data and are widely used to describe complex real-world systems. Probabilistic Graphical Models (PGMs) and Graph Neural Networks (GNNs)...

How do Probabilistic Graphical Models and Graph Neural Networks Look at Network Data? arxiv.org/abs/2506.11869

Saturday Morning Breakfast Cereal - Dress Saturday Morning Breakfast Cereal - Dress

@lordgrilo.bsky.social this is for you

www.smbc-comics.com/comic/dress

I am listening to Valérie on the latest OMG Genomics podcast episode (omgenomics.com/podcast) about the quality and importance of annotation. Basically SNAFU, I hope we will improve the situation, also valuing the curators and their work.

Slightly diminish a band:

U1

When I first started working in omics I used to joke about the fact you can likely find a link between any gene and any mechanism. I’ve tested Biomni co-pilot today with random SNP/phenotype, it’s amazing what it can do and how it masters our joke.
I can’t say if it’s a bad or a good thing

Is single-molecule protein sequencing here yet? - Nature Methods As instruments and approaches emerge for single-molecule protein analysis, some developers and early users share their first impressions.

I love MS, I've started with MS. And yet I can't wait for protein sequencing to come

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

#1 strip

Quality control of single-cell ATAC-seq data without peak calling using Chromap In this work, we extend Chromap, an ultrafast method for single-cell ATAC-seq data alignment, to directly report peak-based quality control (QC) metrics, such as the fraction of reads in peaks, withou...

Cool, I’ve missed this new QC feature!

www.biorxiv.org/content/10.1...

BTW, after reading the paper it seems to me they took their previous work (VeloVAE) and applied to a spatial graph. I wonder how other velocity methods perform if used in the same way (e.g. scVelo smoothed on spatial graph instead of/in addition to kNN graph)

Well, of course mine was a joke, at least the tissue slide has an obvious physical interpretation :-)

Topological velocity inference from spatial transcriptomic data - Nature Biotechnology TopoVelo infers cell differentiation stages and rate parameters of gene expression using spatially coupled differential equations.

Haven’t read this yet, but it seems that the good old embedding problem is solved 😅

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

As said some time ago, I will

Too bad I won’t be there! Looking forward for the GR paper, then…