Understanding how genetic regulatory variation affects gene expression in different T cell states is essential to deciphering autoimmunity. We conducted a high-resolution RNA-seq time course analysis of stimulated memory CD4 + T cells from 24 healthy individuals. We identified 186 genes with dynamic allele-specific expression, where the balance of alleles changes over time. These genes were four fold enriched in autoimmune loci. We found pervasive dynamic regulatory effects within six HLA genes, particularly for a major autoimmune risk gene, HLA-DQB1. Each HLA-DQB1 allele had one of three distinct transcriptional regulatory programs. Using CRISPR/Cas9 genomic editing we demonstrated that a single nucleotide variant at the promoter is causal for T cell-specific control of HLA-DQB1 expression. Our study in CD4 + T cells shows that genetic variation in cis regulatory elements may affect gene expression in a lymphocyte activation status-dependent manner contributing to the inter-individual complexity of immune responses.Genetic studies have identified an enrichment of autoimmune risk alleles in memory CD4 + T cell-specific regulatory elements(1-3). Memory CD4 + T cells are essential orchestrators of immune response. Hence, it is crucial to study how genetic variation affects their gene expression patterns to unravel the complex dynamics of regulation. Previous studies on activated T cells analyzed a limited number of cell states and genes(4-9), and an understanding of how gene expression levels are influenced by genetic regulatory variation in multiple physiological states is lacking. In this study, we investigated activation-dependent genetic regulatory effects in memory CD4 + T cells by studying dynamic allele-specific expression that changes with time in a high resolution RNA sequencing time series.Studying allele-specific expression (ASE) of genes can enable the detection and characterization of context-specific cis regulatory effects (10,11). In a pilot experiment, we stimulated memory CD4 + T cells from two genotyped individuals of European ancestry ( fig. S1) with anti-CD3/CD28 beads. We ascertained gene expression at 0, 2, 4, 8, 12, 24, 48 and 72 hours after stimulation using deep mRNA sequencing (Fig. 1A). Using a logistic regression framework, we identified dynamic ASE (dynASE) events (Methods) at heterozygous SNPs. These dynASE sites are those where the imbalance of the two expressed alleles is time dependent. First, for each heterozygous site in an individual, we merged counts from all time points and identified 1,484 sites with evidence of significant ASE (intercept P < 2.8x10 -6 =0.05/17,743 tests, Bonferroni threshold). Next, for those sites we assessed time-dependent ASE effects by fitting a second order polynomial model. To account for over-dispersion of allelic counts, we incorporated sample-to-sample variability with a random intercepts effect. We observed 64 dynASE events in these two individuals (P < 3.7e-03, <5% FDR) in 60 SNPs, in 37 genes. In an independent experiment for the same two individu...