Study Questions Effectiveness of Foundation Models for Genomics Due to High Entropy in DNA Sequences
Researchers analyzing foundation models trained on genomic data found that DNA sequences have inherently high entropy, causing models to produce near-uniform predictions and unstable outputs. Foundation models have underperformed in genomics compared to natural language processing, but the underlying reasons were unclear until this investigation. The findings suggest that current self-supervised training approaches may be fundamentally unsuitable for genomic data, potentially requiring new methodologies.
A new study published on arXiv examines why foundation models—large neural networks trained on broad datasets—have achieved limited success in genomics despite their effectiveness in natural language processing. Researchers trained ensembles of models on both text and DNA sequences, then analyzed their predictions, embeddings, and information flow patterns. They discovered that genomic sequences exhibit high entropy from the perspective of unseen token prediction, leading to near-uniform output distributions, disagreement between identically-trained models, and unstable static embeddings. The analysis of Fisher information revealed that models trained on DNA concentrate information in embedding layers rather than exploiting relationships between tokens. These findings challenge the foundational assumptions of current genomic foundation model training, suggesting that self-supervised learning from sequences alone may not be applicable to genomic data and that alternative approaches may be necessary.
What's missing
The study's own limitations and open questions are not detailed in the abstract provided. Specifically, it is unclear whether the findings apply uniformly across different types of genomic sequences, whether alternative training objectives or architectures might overcome the entropy challenge, or what specific new methodologies the authors propose as solutions.
What different sources said
- arXiv cs.LGCenter
Entropy, Disagreement, and the Limits of Foundation Models in Genomics
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.