Explicit statistical models convert bulk and single-cell omics into inferential claims

Published papers

• Decoding cancer circulating transcriptomic signatures with language models.

  Nature communications2026-07-09Siwei Deng et al.

  GeneLLM, a Transformer that models nucleotide sequences of human-mapped plasma cfRNA reads without gene quantification, reports ROC-AUC 0.9250-0.9962 for…

• Statescope: an integrative deconvolution framework for discovering cell states in tumors.

  Nature communications2026-07-06Jurriaan Janssen et al.

  Statescope is a Bayesian deconvolution framework that uses DNA-derived malignant purity to infer cell states, including states absent from single-cell…

• Generalizable AI predicts immunotherapy outcomes across cancers and treatments.

  Nature medicine2026-07-03Wanxiang Shen et al.

  COMPASS, a pan-cancer foundation model with a concept bottleneck transformer, predicts immunotherapy response from bulk tumor transcriptomes via 44 immune…

• Deep learning models for cell cycle phase prediction from single-cell RNA sequencing data.

  Briefings in bioinformatics2026-07-03Halima Akhter et al.

  Benchmarks traditional and deep learning models for predicting cell cycle phase from scRNA-seq, trained on consensus labels and tested on independent human and…

• Exploring per-base quality scores as a surrogate marker of cell-free DNA fragmentome.

  Briefings in bioinformatics2026-07-03Hadas Volkov et al.

  Per-base quality scores as alignment-free cfDNA fragmentomic surrogate; leave-one-batch-out AUC 0.81 separating cancer from controls.

• MRDsteer: quality-aware AI-driven closed-loop optimization enhances ctDNA-based minimal residual disease detection.

  Briefings in bioinformatics2026-07-03Tianci Wang et al.

  MRDsteer, an AI-driven closed-loop agent, monitors variant-calling reliability in ctDNA analysis and triggers targeted re-calling, improving ultra-low-VAF MRD…

• miRSiC: a regulatory-aware machine learning framework for microRNA expression inference across bulk and single-cell transcriptomes.

  Briefings in bioinformatics2026-07-03Guan-Ting Chen et al.

  miRSiC is an interpretable LightGBM framework inferring miRNA expression from TF and validated target gene features, tested on TCGA breast cancer and…

Preprints

• SHIFT: Survival Prediction from Incomplete and Heterogeneous Genomic Data

  arXiv2026-07-04Muhammet Sami Yavuz et al.

  SHIFT is a missingness-aware transformer for survival prediction from incomplete genomic panels, using masked self-attention and variable-rate feature masking,…

• Computing tumor specificity of cancer antigen targets by k-mer indexing of healthy tissue transcriptomes

  bioRxiv2026-07-03Hausmann, J. et al.

  k4neo uses k-mer indexing of raw RNA-seq reads across 18,960 healthy-tissue samples to compute tumor specificity of transcript variants for cancer antigen…

• CellTFusion: a transcriptional regulatory network framework for the identification of functional multicellular states from bulk RNA-seq data

  bioRxiv2026-07-05Hurtado, M. et al.

  CellTFusion integrates cell type deconvolution with transcriptional regulatory network analysis to identify functional multicellular TME states from bulk…

Related topics

• Machine Learning Tumor Profiling
• Computational Cancer Genomics Inference
• Single Cell Clinical Outcome Prediction

Topic history