Paper available here; Model available at https://ploginovic.shinyapps.io/ms_grs_predict/ and mspredictor.com;
This repository contains scripts used to generate MS-GRS and assess it in the context of undifferentiated Optic Neuritis in UKBB, replicated in FinnGen (Finland) and Geisinger-Regeneron (DiscovEHR)
HLA-GRS was created using direct_hla_interaction_score.ipynb and 2 SNPs scored in PLINK2 using plink_twoHLA_snps.sh. Interaction HLA GRS is based on a 10-allele model reported by Moutsianas et al (2015, supplement page 15)
non-HLA GRS was created using file nonhla_plink_2201.sh
non-HLA and HLA-GRS were merged using demo_nonhla_merge_2201.r, which also produced the final .csv file used in Stata analysis later
phenotypic analysis was done in stata using on_ms_ukb_v2_1.do
Figures and subsequent ROC and survival analysis were performed in Python final_step_analysis.ipynb
Moutsianas, L., Jostins, L., Beecham, A.H., Dilthey, A.T., Xifara, D.K., Ban, M., Shah, T.S., Patsopoulos, N.A., Alfredsson, L., Anderson, C.A. and Attfield, K.E., 2015. Class II HLA interactions modulate genetic risk for multiple sclerosis. Nature genetics, 47(10), p.1107.