First Study Demonstrates Scaling Laws for Transformer Models in Single-Cell Genomics
Researchers conducted the first systematic study of neural scaling laws for masked-reconstruction transformers trained on single-cell RNA sequencing data, finding that clear power-law scaling emerges when sufficient data are available. The study used expression profiles from the CELLxGENE Census across seven model sizes ranging from 533 to 3.4×10⁸ parameters, revealing that the data-to-parameter ratio is critical to scaling behavior. These findings could inform the design of foundation models for genomics and establish a framework for understanding model performance in single-cell transcriptomics.
This arXiv preprint presents the first systematic investigation of scaling laws—power-law relationships between loss, model size, and data—in single-cell genomics, extending principles previously documented in language and vision models. The researchers trained masked-reconstruction transformers on single-cell RNA sequencing (scRNA-seq) data from the CELLxGENE Census under two conditions: a data-rich regime (512 genes, 200,000 cells) and a data-limited regime (1,024 genes, 10,000 cells). In the data-rich regime, they observed clear power-law scaling with an irreducible loss floor of approximately 1.44, while the data-limited regime showed negligible scaling, indicating that data scarcity—not model capacity—is the limiting factor. The researchers converted their findings to information-theoretic units, estimating approximately 2.30 bits of entropy per masked gene position. The work establishes that scaling laws analogous to those in natural language processing do apply to single-cell transcriptomics under appropriate data conditions and identifies the data-to-parameter ratio as a key determinant of model performance.
What's missing
The study acknowledges that additional measurements are needed to refine the entropy estimate and does not provide direct validation on downstream biological tasks (e.g., cell type prediction, disease classification) that would demonstrate practical utility of the scaling laws for genomics applications.
What different sources said
- arXiv cs.LGCenter
Scaling Laws for Masked-Reconstruction Transformers on Single-Cell Transcriptomics
Related
Gut Bacteria Enzyme Found to Break Down Heat-Processed Food Compounds, Producing Novel Biogenic Amines
Researchers have discovered that an enzyme in common gut bacteria can degrade N-epsilon-carboxymethyllysine (CML), a compound formed during thermal food processing, producing previously unknown biogenic amines. The enzyme, ornithine decarboxylase SpeC from enterobacteria, acts on CML and related modified lysine derivatives through a low-level 'underground' catalytic activity. This finding suggests a previously unrecognized communication axis between thermally processed dietary compounds and gut microbial physiology, with potential implications for host health.
Full-Length Gene Sequencing Reveals Two Distinct Bacterial Communities in Black-Legged Ticks Expanding Into Canada
Researchers used Oxford Nanopore full-length 16S rRNA gene sequencing to characterize the microbiome of Ixodes scapularis black-legged ticks collected in Nova Scotia, Canada, distinguishing between tick-adapted bacteria and environmentally acquired bacteria. The study comes as I. scapularis — the primary vector of Lyme disease — is rapidly expanding northward into Canada due to climate change. The findings suggest that environmentally derived bacteria in tick microbiomes are not mere contamination, which has implications for how tick microbiome data is collected and interpreted across surveillance studies.
Study Identifies Metabolic Link Between Cell Envelope Stress and Biofilm Formation in Bacteria
Researchers have discovered that the metabolite acetyl-CoA directly inhibits enzymes that degrade the bacterial signaling molecule c-di-GMP, connecting cell envelope biosynthesis stress to biofilm formation in Pseudomonas aeruginosa. The study found that sub-inhibitory concentrations of antibiotics targeting early peptidoglycan biosynthesis — but not other antibiotic classes — elevate c-di-GMP levels by reducing phosphodiesterase activity, with acetyl-CoA competing for the enzyme active site. Because the relevant enzyme domain is broadly conserved across bacterial species, this checkpoint mechanism may be widespread and could have implications for understanding antibiotic-induced biofilm responses.