Research Letter: Therapeutic targets for haemorrhoidal disease: proteome-wide Mendelian randomisation and colocalization analyses

Abstract

Human haemorrhoidal disease (HEM) is a common anorectal pathology. However, being one of the diseases that affect a wide range of people, the etiology of HEM, as well as its molecular mechanism, remains primarily unclear. In this study, we applied a two-sample bi-direction Mendelian randomization (MR) framework to estimate the causal effects of 4677 plasma proteins on HEM outcomes and investigated the mediating impacts of plasma proteins on HEM risk factors to uncover potential HEM treatment targets by integrating GWASs statistics of HEM and plasma protein levels. Following MR analysis, our study identified 13 probable causal proteins associated with HEM. Particularly, genetically predicted OLFM1 levels were linked to an increased risk of HEM. In addition, it was discovered that a genetically greater risk of myxoedema, diverticular disease, and ulcerative colitis was linked to an elevated risk of HEM (FDR<0.05). However, there was no evidence that dorsalgia, hernia, and ankylosing spondylitis were causally associated with HEM. Interestingly, a higher risk of myxoedema was associated with higher genetically predicted OLFM1 levels. Finally, mediation analysis suggested that the effect of myxoedema on HEM via OLFM1 might explain 32.8% of the mediation effect. Overall, this study identified some causal associations of circulating proteins and risk factors with HEM by integrating the largest-to-date plasma proteome and GWASs of HEM. The findings could provide further insight into understanding biological mechanisms for HEM.

Identification of therapeutic targets for 18 clinical diseases by integrating human plasma proteome with genome

“…Single-cell RNA-sequencing of 54,762 cells taken from normal thyroid tissue from 7 individuals who had thyroidectomy and had verified instances of differentiated thyroid carcinoma were retrieved from Gene Expression Omnibus (https://www.ncbi.nlm.nih.gov/geo/, GSE182416) [16]. Seurat 4.2.0 was used for processing and analyzing the single-cell dataset, as previously stated [13]. In short, low-quality cells by selecting feature numbers ranging from 200 to 3,000 were excluded.…”

“…Based on the presence of non-overlapping samples between exposure (4,907 proteome data) and outcome (16 clinical traits), two-sample MR analysis has been performed using the TwoSampleMR, as previously described [13][14]. In brief, the data were harmonized after clumping (r 2 <0.001) to exclude ambiguous SNPs with non-concordant alleles and palindromic SNPs.…”