Fig. 2: Association of germline DDX41 variants with disease contexts and ancestral groups.

A The proportion (%) of cases with pathogenic/likely pathogenic (P/LP) or uncertain (VUS) DDX41 variants across various disease contexts in the literature cohorts. Whiskers indicate the 95% confidence interval. Twelve cases had two germline variants (6 with both VUS and P/LP were counted twice), and 7 cases had a lymphoid neoplasm in addition to either MDS/AML (n = 6) or another myeloid neoplasm (n = 1) (counted twice). Pairwise Fisher’s exact tests were adjusted using the Benjamini-Hochberg method. P-value annotations: <0.05 (*), < 0.01 (**), < 0.001 (***), < 0.0001 (****). B The stacked bar chart illustrates the distribution of 1374 germline DDX41 variants in 1366 MDS/AML cases across different classes of pathogenicity. Two cases were unclassified due to missing HGVSc information and were excluded. Ten cases harbored two germline variants: 6 were VUS + P/LP, 3 were both VUS, and 1 was both LP. Variant counts for each classification are shown below each bar. C Quasi-case-control analysis of germline DDX41 variants in patients with MDS/AML compared to population controls, differentiated by overall and ancestral groups (non-Finnish European [NFE] versus East Asian [EAS]). Variants with at least 10 total cases, 3 occurrences within the EAS ancestry group, and 5 NFE-only instances are included. The odds ratio and its 95% confidence interval are shown. Gray, blue, and red represent overall, NFE, and EAS ancestry groups, respectively, based on the exclusivity (or lack thereof) of variants within each ancestry group and the control group. Red circles to the left of the variant indicate downgraded variants due to added ancestry matching: PS4_moderate (two solid circles), PS4_supporting (single solid circle), or PS4_notmet (open red circle).