Some useful links are as follows:
This website layout was inspired by the layout of the CAUSE website.
Studying genetic associations with prognosis (e.g., survival, disability, subsequent disease events) or conditional on a phenotype (e.g., disease incidence) may be affected by selection bias - also termed index event bias or collider bias - whereby selection on disease status can induce associations between causes of incidence with prognosis.
The ‘Slope-Hunter’ approach is proposed for adjusting genetic associations for this bias. The approach is unbiased even when there is genetic correlation between incidence and prognosis.
Our approach uses advanced machine learning techniques such as unsupervised model-based clustering tailored to theoretical distributions of genetic effects on incidence and prognosis. The ‘Slope-Hunter’ method identifies and utilises the cluster of genetic variants only affecting incidence to estimate an unbiased adjustment factor for collider bias correction even in the presence genetic correlations (e.g., shared genetic pathways as typically observed for many traits including metabolites, cancer risk factors, psychiatric phenotype). The ‘Slope-Hunter’ approach assumes the identified cluster of variants only affecting incidence explains more variation in incidence than any other variant clusters.
An easy-to-follow practical tutorial on how to implenment the method
To Appear: Gallery of corrected prognosis GWAS (e.g. disease progression). If you are interested to deposit your prognosis GWAS data after correction in the Slope-Hunter gallery, please do contact Dr. Osama Mahmoud on o.mahmoud@essex.ac.uk
If you use the ‘Slope-Hunter’ method, please do consider to cite both the paper and the software as follows:
Note: in the citation below, replace the
Add the following code to your website.
For more information on customizing the embed code, read Embedding Snippets.