Prediction of treatment response in rheumatoid arthritis patients using genome‐wide SNP data

Abstract

Although a number of treatments are available for rheumatoid arthritis (RA), each of them shows a significant nonresponse rate in patients. Therefore, predicting a priori the likelihood of treatment response would be of great patient benefit. Here, we conducted a comparison of a variety of statistical methods for predicting three measures of treatment response, between baseline and 3 or 6 months, using genome‐wide SNP data from RA patients available from the MAximising Therapeutic Utility in Rheumatoid Arthritis (MATURA) consortium. Two different treatments and 11 different statistical methods were evaluated. We used 10‐fold cross validation to assess predictive performance, with nested 10‐fold cross validation used to tune the model hyperparameters when required. Overall, we found that SNPs added very little prediction information to that obtained using clinical characteristics only, such as baseline trait value. This observation can be explained by the lack of strong genetic effects and the relatively small sample sizes available; in analysis of simulated and real data, with larger effects and/or larger sample sizes, prediction performance was much improved. Overall, methods that were consistent with the genetic architecture of the trait were able to achieve better predictive ability than methods that were not. For treatment response in RA, methods that assumed a complex underlying genetic architecture achieved slightly better prediction performance than methods that assumed a simplified genetic architecture.

Machine learning for effectively avoiding overfitting is a crucial strategy for the genetic prediction of polygenic psychiatric phenotypes

“…When the prediction accuracies of all conducted models were calculated using the independent validation cohort, the STMGP prediction model showed the highest partial correlation (Table 2), but it was not significantly different from that of the other prediction models (P values > 0.05). When the training Since ridge regression based on raw SNP data was difficult to implement in our environment due to the substantial computational cost, the genome data were clumped into approximately 30,000 SNPs in a manner similar to a previous study for these analyses 51 .…”

An evaluation of machine-learning for predicting phenotype: studies in yeast, rice, and wheat

“…Standard off-the-shelf machine learning methods (Dudoit et al 2002;Ziegler et al 2007;Szymczak et al 2009;Ogutu et al 2011Ogutu et al , 2012Pirooznia et al 2012;Mittag et al 2012;Okser et al 2014;Leung et al 2016;Cherlin et al 2018) present an attractive alternative to classical statistical genetics methods. These methods are easy to use, are freely available in a variety of implementations, and intrinsically multivariate.…”

Association of response to TNF inhibitors in rheumatoid arthritis with quantitative trait loci for CD40 and CD39

“…Genome-wide association studies (GWAS) of response to TNFi have shown that common single nucleotide polymorphisms (SNPs) explain an estimated 40% and 50% of the variance of change in swollen joint counts (SJC) and erythrocyte sedimentation rate (ESR), respectively; however, no strong associations with individual SNPs have been detected 4. Thus, as with many complex phenotypes, the genetic architecture of response to TNFi is likely to be polygenic with many small genetic effects 5. In this situation, the sample size required to learn a predictive model is very large—up to 10 cases per variable6—and it may not be feasible to assemble such large sample sizes for studying response to a single drug or drug class.…”