Figure 2.

Distribution of γ-retroviral vector integration sites (ISs) relative to genomic features. (a) Distances from ISs to the nearest “UCSC Known Gene” transcription start site (TSS) were computed, binned in 10-bp windows, converted to proportions relative to the total number of sites in the respective datasets, and are plotted for those sites falling within 2.5 kb either side of the nearest TSS. MRCs, matched random control sites. (b) Integration site distributions relative to coding genes. Percentages of sites within each genic category are shown in the columns, and the total number of sites per dataset is indicated below. Comparison of IS distributions between P and L datasets shows very strong statistical evidence of differences (χ² = 473, df = 2, P < 0.0001). The proportions of intragenic integration do not differ between the two datasets (two-tailed Fisher’s exact test on IS counts, P = 0.26). The P values for all comparisons within genic categories are presented in Supplementary Tables S3–S5. (c) The extent of association between annotated genomic features and vector ISs is summarized in the form of heat maps. Columns are labeled with IS dataset names and rows with analyzed genomic features. The colored receiver operating characteristic (ROC) curve area scale to the right of the panel shows increased integration (relative to MRC sites) near the indicated feature in red, decreased integration in blue, with the intensity of shading correlating with the degree of departure from random integration. Detailed comparisons of relative integration abundance associated with individual genomic features are available as “Supplementary report - Association of Genomic Features with Integration”.