Exploring the genetic architecture of multiple long-term conditions using a genome-wide association study in the UK...

The prevalence of multiple long-term conditions (MLTC) is increasing. It is essential to develop strategies to prevent and manage MLTC; however, the biological mechanisms underlying MLTC are not yet clearly understood. We used UK Biobank data as part of the ADMISSION research collaborative to identify genetic drivers for MLTC. We used the UK Biobank (UKBB) self-reported illness data to characterise MLTC (defined as two or more long-term conditions) using 51 common disease labels. A genome-wide association study (GWAS) was conducted for MLTC and complex MLTC (complex MLTC was defined as having three or more diseases from the 51 self-reported diseases, with these three diseases additionally belonging to different body systems), and post-GWAS analyses were conducted to explore the genomic loci associated with MLTC. We then undertook a factor analysis on the individual-level disease data to identify the factors contributing to MLTC. We investigated the genomics of these factors using single disease polygenic risk score (PRS) and GWAS. The prevalence of simple MLTC was 33.0% (n = 111,184) and complex MLTC was 11.2% (n = 37,650). The majority (81.3%) of significant SNPs from MLTC GWAS were located in chromosome 6 with most of them in the HLA region. The 'T cell activation' pathway and apoptosis signalling pathways were identified in gene-based pathway analysis. Five latent factors were identified through factor analysis with the following underlying characteristics: Factor 1, metabolic disease; Factor 2, mental ill health; Factor 3, cancer; Factor 4, musculoskeletal and inflammation-related traits; Factor 5, digestive system-related diseases. The GWAS and PRS-based analysis validated the characteristics of these factors. The MLTC GWAS, complex MLTC GWAS and factor-based GWAS analyses highlighted the association between HLA genes and MLTC. Further research is needed to disentangle the association between MLTC and the HLA genes, along with the integration of multi-omics data.