Disease heterogeneity dissection

Tue, 01 Jul 2025 00:00:00 +0000

Diagnosis for the same disease affects individuals differently. However, we don’t know how best to represent biologically meaningful axes of variation within a disease, beyond the standard case vs. control comparison or proxy measure of severity. We combine human genetics, causal inference, and modern AI to address this challenge.

Decomposition of genetic associations across many phenotypes
Causal representation learning from pleiotropic associations
Assessing the biological bases of disease subtypes

Representative studies

UK Biobank Latent Components Obesity +1 more

Components of genetic associations across 2,138 phenotypes in the UK Biobank highlight adipocyte biology

We developed DeGAs to decompose shared genetic associations across 2,138 UK Biobank phenotypes and identify their underlying pleiotropic structure.

Yosuke Tanigawa, Ph.D.

• Dec 31, 2019 • 1 min read

Polygenic Scores Latent Components Disease Heterogeneity

Polygenic risk modeling with latent trait-related genetic components

Polygenic risk score (PRS) has been proposed for disease risk prediction with potential clinical relevance for some traits, but its personalized interpretation is generally …

Feb 8, 2021 • 1 min read

UK Biobank Phenotyping Latent Components +1 more

A cross-population atlas of genetic associations for 220 human phenotypes

Using a set of GWAS summary statistics of diseases characterized from both European (UK Biobank and FinnGen) and East Asian (Biobank Japan) populations, we dissected latent DeGAs …

Sep 30, 2021 • 1 min read

Predictive modeling of disease

Thu, 11 Jan 2024 00:00:00 +0000

Accurate prediction of disease risk using genomics enables earlier identification of individuals at elevated risk and facilitates more effective prevention strategies. However, existing genetic approaches show limited generalizability across cohorts and lack biological interpretation. We develop robust and biologically interpretable models using high-dimensional data.

Systematic integration of genomic annotations into predictive models
Biologically interpretable sparse models for genetic prediction of disease
Flexible predictive modeling applicable across distinct contexts

Representative studies

Polygenic Scores Sparse Models Epigenomics +1 more

PRISM: ancestry-aware integration of tissue-specific genomic annotations enhances the transferability of polygenic scores

In this study, led by Xiaohe (Lucy) Tian, we showed that ancestry-aware integration of tissue-specific genomic annotations enhances the transferability of polygenic scores (PGS).

Lucy Tian

• Nov 13, 2025 • 1 min read

Polygenic Scores Resources Sparse Models +1 more

Power of inclusion: Enhancing polygenic prediction with admixed individuals

We developed a polygenic score training approach that allows direct inclusion of admixed individuals without the need for local ancestry inference and showed ancestry-diverse …

Yosuke Tanigawa, Ph.D.

• Oct 26, 2023 • 1 min read

UK Biobank Polygenic Scores Resources +1 more

Significant Sparse Polygenic Risk Scores across 813 traits in UK Biobank

We performed a systematic assessment of the predictive performance of PRS models across >1,500 traits in UK Biobank and report 813 PRS models with significant predictive …

Yosuke Tanigawa, Ph.D.

• Mar 24, 2022 • 1 min read

Statistical Genetics | Tanigawa Lab

Disease heterogeneity dissection

Representative studies

Components of genetic associations across 2,138 phenotypes in the UK Biobank highlight adipocyte biology

Polygenic risk modeling with latent trait-related genetic components

A cross-population atlas of genetic associations for 220 human phenotypes

Predictive modeling of disease

Representative studies

PRISM: ancestry-aware integration of tissue-specific genomic annotations enhances the transferability of polygenic scores

Power of inclusion: Enhancing polygenic prediction with admixed individuals

Significant Sparse Polygenic Risk Scores across 813 traits in UK Biobank