Unconditional, multivariate logistic regression modeling incorporating all detected alleles from sequencing-based classical HLA typing at six loci.

Significant contributions to disease status were established on the basis of a likelihood ratio test of the impact from each allele on the Akaike Information Criterion (P≤0.01). Due to the strong linkage disequilibrium (LD) between DRB1 and DQB1 alleles, and to reduce the degree of freedom of the test, they were modeled as “DRB1-DQB1 blocks” according to LD. aDRB1*13:XX = all non-13:01 alleles, i.e. DRB1*13:02, 13:03 and 13:05.CI: confidence interval, OR: odds ratioUnconditional, multivariate logistic regression modeling incorporating all detected alleles from sequencing-based classical HLA typing at six loci.

本研究基于各等位基因对赤池信息准则（Akaike Information Criterion, AIC）的影响开展似然比检验（likelihood ratio test），据此证实了各等位基因对疾病状态的显著贡献（P≤0.01）。鉴于DRB1与DQB1等位基因间存在强连锁不平衡（linkage disequilibrium, LD），且为降低检验的自由度，研究人员依据连锁不平衡将二者建模为“DRB1-DQB1区块”。其中aDRB1*13:XX指代所有非13:01型等位基因，即DRB1*13:02、13:03及13:05。CI为置信区间（confidence interval, CI），OR为比值比（odds ratio, OR）。本研究采用无条件多变量logistic回归模型，纳入了基于测序的6个基因座经典HLA分型（HLA typing）所检测到的全部等位基因。