scDRS-FM is a method that links scRNA-seq with polygenic risk of disease to fine-map disease associations across individual cells in the scRNA-seq data, as described in our paper “Conditional polygenic enrichment distinguishes causal from tagging disease-critical cell populations in single-cell RNA-seq” (Turcan et al. 2026b).
TusoAI is an agentic system for autonomous method development, catered towards computational biology but applicable to various fields, as described in our paper “TusoAI: Agentic Optimization for Scientific Methods” (Turcan et al. 2026a).
KGWAS is a novel geometric deep learning method that leverages a massive functional knowledge graph across variants and genes to improve detection power in small-cohort GWASs, as described in our paper “Small-cohort GWAS discovery with AI over massive functional genomics knowledge graph” (Huang et al. 2024).
LDSPEC is a method for estimating the correlation of causal disease effect sizes for pairs of nearby SNPs, depending on their functional annotations, as described in our paper “Pervasive correlations between causal disease effects of proximal SNPs vary with functional annotations and implicate stabilizing selection” (Zhang et al. 2023).
scDRS is a method that links scRNA-seq with polygenic risk of disease at individual cell resolution, as described in our paper “Polygenic enrichment distinguishes disease associations of individual cells in single-cell RNA-seq data” (Zhang*, Hou* et al. 2022).
Popular k-medoids clustering algorithms, such as Partitioning Around Medoids (PAM), are prohibitively expensive in computation for large datasets. BanditPAM is a randomized version of PAM that returns the same results with high probability but is substantially faster. The algorithm is described in our paper “BanditPAM: Almost Linear Time k-Medoids Clustering via Multi-Armed Bandits” (Tiwari et al. 2020).
Empirical Bayes estimators for single-cell RNA-seq analysis, as described in our paper “Determining sequencing depth in a single-cell RNA-seq experiment” (Zhang*, Ntranos* et al. 2020).
AdaFDR is a fast and covariate-adaptive method that learns adaptive p-value thresholds from covariates to improve power while controlling FDR. The method is described in our paper “Fast and covariate-adaptive method amplifies detection power in large-scale multiple hypothesis testing” (Zhang et al. 2019).
cPCA is a dimensionality reduction algorithm that identifies low-dimensional structures that are enriched in a dataset relative to comparison data. Applications include dicovering subgroups in biological and medical data. The method is described in our paper “Exploring patterns enriched in a dataset with contrastive principal component analysis” (Abid*, Zhang* et al. 2019).