Fig. 4: Analysis of the longest TRs as disease-associated TRs and characterization of their generating units.

a The unit length distribution of the most frequent units of the longest TRs such that the TR length difference between the longest and median TRs is in the range shown at left. The second left shows the numbers of TR loci. The boxplot inside each violin plot shows the minima, 1st quartile, 2nd quartile (center), 3rd quartile, and maxima. b Analysis of expanded copies of the 69-mer unit at chr18:57024495-57024955 in the human reference genome (hg38) that are correlated with ALS samples of European descent. The red boxplot below the histogram shows the 1st, 2nd, and 3rd quartiles of copy number of the 69-mer unit in the ALS samples. The histogram shows the density distribution of the copy number of the 69-mer unit in Japanese control samples (n = 270). c TR loci are classified according to whether the longest TR representative is complex TR (red) or single-unit TR (blue). Single-unit longest TRs are more than complex ones when the length difference is 100 or more. d The unit length distribution of the most frequent key unit in the longest TR representative of 270 independent individuals. The boxplot inside each violin plot shows the minima, 1st quartile, 2nd quartile (center), 3rd quartile, and maxima. Similar to Fig. 3f, units in complex TRs were significantly shorter than single units (p < 10⁻¹¹ according to two-sided Spearman’s rank test) when the TR length difference was >100 b.