In a paper published in Nature Genetics, Finucane et al. combine genome-wide association study (GWAS) summary statistics with gene expression data to identify disease-relevant cell types and tissues.

Using five gene expression data sets from humans and mice across multiple tissues, the team set out to investigate whether a specific tissue was enriched for a given trait. Linear regression of genome-wide associations of genetic variants with disease phenotypes enabled the group to evaluate the contributions of specific tissues and cell types to trait heritability and disease in a technique they call linkage disequilibrium score regression applied to specifically expressed genes (LDSC-SEG).

Credit: Judith Collins/Alamy

In total, the researchers investigated 48 diseases and traits. For 34 of them, at least 1 tissue showed significant contribution to the trait. This approach helped to confirm known biological asstotalociations; for example, psychiatric traits were associated with brain-specific gene expression.

The use of multiple tissues also enabled the team to look at contributions from different sources to disease. The analysis of immune-related traits and gene expression showed enrichment for variants associated with bipolar disorder and schizophrenia, alongside immunological diseases, but not other psychiatric diseases.

analysing polygenic signals in multiple tissues … was able to identify the contribution of many tissues and cell types to a disease

The group also investigated the association of genetic variants with chromatin data from the Roadmap Epigenomics and ENCODE projects to determine the role that genome architecture may have in disease. The analysis of chromatin data included some cell types that were absent from the gene expression data. In one case this enabled the detection of an enrichment for neurological tissues in a migraine data set including both migraine with and without aura; this enrichment was absent when looking only at migraine without aura, suggesting that different tissues may be relevant for different migraine subtypes.

By analysing polygenic signals in multiple tissues, the group was able to identify the contribution of many tissues and cell types to a disease. While these associations will be difficult to evaluate in vitro and in vivo, as the authors caution, this study helps to lay the groundwork for future investigations into the mechanisms of disease pathology.