Fig. 4: Structural variants associated with AD. | Nature Communications

Fig. 4: Structural variants associated with AD.

From: Whole-genome sequencing analyses suggest novel genetic factors associated with Alzheimer’s disease and a cumulative effects model for risk liability

Fig. 4

a Overview of association test between structural variants and AD. b Manhattan plot for association analysis of CNV with Aβ positivity. Red and gray dashed lines indicate the Bonferroni (p = 2.4 × 10−6) and cytoband-based Bonferroni (p = 1.6 × 10−4) thresholds, respectively. Two-sided logistic regression was used with sample covariates. c Schematic of three STR association models. d Proportion of tandem repeats observed within each genomic region. The proportion reflects the frequency of STR presence relative to the size of the region. Dashed line: genome-wide average. e Histogram of STR motif length (inset: ≥ 7 bp). f Histogram of median STR tract lengths relative to GRCh38, genotyped by ExpansionHunter; x-axis limited to – 20 to 20. g Histogram of STR repeat counts in the GRCh38. h The association between individual STRs (x-axis: mean length difference between Aβ-positive and Aβ-negative samples divided by standard deviation). STR lengths were analyzed using two-sided logistic regression, adjusting for sample, technical, and APOE ε4 covariates. i Relative burden of STRs enriched in Aβ-positive individuals across thresholds of STR length and observation count. j Distribution of STR outlier counts for each sample compared across Aβ-positive (n = 895) and Aβ-negative samples (n = 620). k Odds ratio measuring the likelihood of Aβ positivity in individuals with different numbers of STR outliers. l Aβ levels of each sample were compared between groups and divided according to the threshold ( < 10 or ≥ 40) for the STR outlier count. m Same as (l), but for CU samples only. n Gene set enrichment analysis for genes near STR outliers in samples with ≥ 40 outliers (two-sided Fisher’s exact test, FDR-adjusted). The box plots in this figure show the median line. Box edges mark the 25th and 75th percentiles. Whiskers span 1.5 × the interquartile range. Points beyond are outliers. For (j, l, m) two-sided linear regression models were used with sample and technical covariates, and n refers to the number of individual samples. Source data are provided as a Source Data file. Aβ, amyloid beta; AD, Alzheimer’s disease; CU, cognitively unimpaired; CNV, copy number variants; STR, short tandem repeat.

Back to article page