Padyukov, Leonid

Objective To test the ability of machine learning (ML) approaches with clinical and genomic biomarkers to predict methotrexate treatment response in patients with early rheumatoid arthritis (RA). Methods Demographic, clinical, and genomic data from 643 patients of European ancestry with early RA (mean age 54 years; 70% female) subdivided into a training (n = 336) and validation cohort (n = 307) were used. The genomic data comprised 160 single‐nucleotide polymorphisms (SNPs) previously associated with RA or methotrexate metabolism. Response to methotrexate monotherapy was defined as good or moderate by the European Alliance of Associations for Rheumatology (EULAR) response criteria at the 3‐month follow‐up. Supervised ML methods were trained with 5 repeats and 10‐fold cross‐validation using the training cohort. Prediction performance was validated in the independent validation cohort. Results Supervised ML methods combining age, sex, smoking, rheumatoid factor, baseline Disease Activity Score in 28 joints (DAS28) scores and 160 SNPs predicted EULAR response at 3 months with the area under the receiver operating curve of 0.84 ( P = 0.05) in the training cohort and achieved a prediction accuracy of 76% ( P = 0.05) in the validation cohort (sensitivity 72%, specificity 77%). Intergenic SNPs rs12446816, rs13385025, rs113798271, and ATIC (rs2372536) had variable importance above 60.0 and along with baseline DAS28 scores were among the top predictors of methotrexate response. Conclusion Pharmacogenomic biomarkers combined with baseline DAS28 scores can be useful in predicting response to methotrexate in patients with early RA. Applying ML to predict treatment response holds promise for guiding effective RA treatment choices, including timely escalation of RA therapies.

Abstract Eating disorders and substance use disorders frequently co‐occur. Twin studies reveal shared genetic variance between liabilities to eating disorders and substance use, with the strongest associations between symptoms of bulimia nervosa and problem alcohol use (genetic correlation [ r g ], twin‐based = 0.23‐0.53). We estimated the genetic correlation between eating disorder and substance use and disorder phenotypes using data from genome‐wide association studies (GWAS). Four eating disorder phenotypes (anorexia nervosa [AN], AN with binge eating, AN without binge eating, and a bulimia nervosa factor score), and eight substance‐use‐related phenotypes (drinks per week, alcohol use disorder [AUD], smoking initiation, current smoking, cigarettes per day, nicotine dependence, cannabis initiation, and cannabis use disorder) from eight studies were included. Significant genetic correlations were adjusted for variants associated with major depressive disorder and schizophrenia. Total study sample sizes per phenotype ranged from ~2400 to ~537 000 individuals. We used linkage disequilibrium score regression to calculate single nucleotide polymorphism‐based genetic correlations between eating disorder‐ and substance‐use‐related phenotypes. Significant positive genetic associations emerged between AUD and AN ( r g = 0.18; false discovery rate q = 0.0006), cannabis initiation and AN ( r g = 0.23; q < 0.0001), and cannabis initiation and AN with binge eating ( r g = 0.27; q = 0.0016). Conversely, significant negative genetic correlations were observed between three nondiagnostic smoking phenotypes (smoking initiation, current smoking, and cigarettes per day) and AN without binge eating ( r gs = −0.19 to −0.23; qs < 0.04). The genetic correlation between AUD and AN was no longer significant after co‐varying for major depressive disorder loci. The patterns of association between eating disorder‐ and substance‐use‐related phenotypes highlights the potentially complex and substance‐specific relationships among these behaviors.

AbstractEpidemiologic studies show an increased risk of non‐Hodgkin lymphoma (NHL) in patients with autoimmune disease (AD), due to a combination of shared environmental factors and/or genetic factors, or a causative cascade: chronic inflammation/antigen‐stimulation in one disease leads to another. Here we assess shared genetic risk in genome‐wide‐association‐studies (GWAS).Secondary analysis of GWAS of NHL subtypes (chronic lymphocytic leukemia, diffuse large B‐cell lymphoma, follicular lymphoma, and marginal zone lymphoma) and ADs (rheumatoid arthritis, systemic lupus erythematosus, and multiple sclerosis). Shared genetic risk was assessed by (a) description of regional genetic of overlap, (b) polygenic risk score (PRS), (c)"diseasome", (d)meta‐analysis.Descriptive analysis revealed few shared genetic factors between each AD and each NHL subtype. The PRS of ADs were not increased in NHL patients (nor vice versa). In the diseasome, NHLs shared more genetic etiology with ADs than solid cancers (p = .0041). A meta‐analysis (combing AD with NHL) implicated genes of apoptosis and telomere length.This GWAS‐based analysis four NHL subtypes and three ADs revealed few weakly‐associated shared loci, explaining little total risk. This suggests common genetic variation, as assessed by GWAS in these sample sizes, may not be the primary explanation for the link between these ADs and NHLs.

ObjectiveTo identify new genetic associations with juvenile and adult dermatomyositis (DM).MethodsWe performed a genome‐wide association study (GWAS) of adult and juvenile DM patients of European ancestry (n = 1,178) and controls (n = 4,724). To assess genetic overlap with other autoimmune disorders, we examined whether 141 single‐nucleotide polymorphisms (SNPs) outside the major histocompatibility complex (MHC) locus, and previously associated with autoimmune diseases, predispose to DM.ResultsCompared to controls, patients with DM had a strong signal in the MHC region consisting of GWAS‐level significance (P < 5 × 10–8) at 80 genotyped SNPs. An analysis of 141 non‐MHC SNPs previously associated with autoimmune diseases showed that 3 SNPs linked with 3 genes were associated with DM, with a false discovery rate (FDR) of <0.05. These genes were phospholipase C–like 1 (PLCL1; rs6738825, FDR = 0.00089), B lymphoid tyrosine kinase (BLK; rs2736340, FDR = 0.0031), and chemokine (C‐C motif) ligand 21 (CCL21; rs951005, FDR = 0.0076). None of these genes was previously reported to be associated with DM.ConclusionOur findings confirm the MHC as the major genetic region associated with DM and indicate that DM shares non‐MHC genetic features with other autoimmune diseases, suggesting the presence of additional novel risk loci. This first identification of autoimmune disease genetic predispositions shared with DM may lead to enhanced understanding of pathogenesis and novel diagnostic and therapeutic approaches.