TifBERT: New Foundation Model for Bulk RNA-Seq Analysis Shows Strong Performance Across Cancer Types
Researchers have developed TifBERT, a self-supervised transformer model that learns representations of bulk RNA sequencing data across normalization schemes and cancer types without requiring expression discretization or external biological embeddings. Pretrained on Pan-Cancer TCGA data spanning five normalization schemes and roughly 10,000 genes, the model achieved 90.83% accuracy and a macro AUC-ROC of 0.996 across 33 cancer types. The work addresses a gap in genomic AI, where foundation models have largely focused on single-cell rather than bulk RNA-seq data, and demonstrates generalization to healthy tissue data from GTEx without retraining.
TifBERT is a self-supervised foundation model designed for bulk RNA sequencing (RNA-seq) representation learning, introduced to address limitations of existing transformer-based approaches that rely on expression discretization, landmark-gene restrictions, or external gene embeddings. The model converts each unordered gene expression profile into a sample-specific gene sequence using TF-IDF ordering — prioritizing genes that are both highly expressed within a sample and selectively expressed across the cohort — and is then pretrained via masked gene modeling, predicting gene identities from transcriptomic context rather than reconstructing raw expression values. Trained on harmonized TCGA Pan-Cancer data across five RNA-seq normalization schemes, TifBERT covers approximately 10,000 genes and achieved 90.83% accuracy, a macro AUC-ROC of 0.996, and a Matthews Correlation Coefficient of 0.903 across 33 cancer types. The model also demonstrated biological interpretability by capturing pathway-level structure, with mean sample-wise and pathway-wise Pearson correlations of 0.754 and 0.762 across 1,387 PARADIGM pathway activities. Notably, TifBERT generalized to GTEx healthy tissue data without retraining, suggesting its representations preserve tissue-level transcriptomic structure beyond the cancer domain. Compared to existing models, TifBERT showed substantially richer embedding geometry, with an effective rank of 95.6 versus 6.3 for competing approaches, alongside greater stability across normalization schemes. The authors position TifBERT as a scalable, normalization-independent framework for reusable bulk transcriptomic representation learning in translational genomics.
What's missing
As a preprint, TifBERT has not yet undergone peer review, and independent replication of its benchmarks has not been reported. The study does not extensively address potential limitations in cohort diversity beyond TCGA and GTEx, nor does it discuss computational resource requirements for pretraining or fine-tuning at scale. The generalizability of TF-IDF ordering to non-cancer clinical datasets or non-human transcriptomes remains untested. The practical downstream utility for clinical prediction tasks beyond cancer type classification is not demonstrated.
What different sources said
- bioRxivCenter
TifBERT: a self-supervised foundation model for normalization-robust bulk RNA-seq representation learning
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.