Fig. 5

Geographical sub-groups in functional segments. a Cluster result of cultivars from 4 regions. The value in heatmap is the median feature-sample score of cultivars from each region. Four regions were clustered into two clades (red and blue dashed lines represent regions where segments belong to sub-group1 and sub-group2 enriched, respectively) by hierarchical method. b Disease index of cultivars from 4 regions in Wang_2017 cohort (n = 195). c t-test for disease index of cultivars from every 2 regions. d The left pieplots represented for ratio of favorable/deleterious segments in SG1 and SG2. The middle scatter plot was the Pearson correlationship between SG1/SG2 segments ratio and disease index (the disease index was normalized by median disease index). The yellow barplot was the SG1/SG2 segment ratio for each cultivar, while the blue barplot was the disease index for each cultivar normalized by median disease index in Wang_2017 cohort (n = 195). e Statistic about number of segments of 4 categories in SG1 and SG2. f The present/absent information of CRI12-specific segments in cultivars from Yangtze River region and the other 3 regions (Yellow River region, Northwestern region, early cotton producing region). g Illustration for deletions of GhKHCP in CRI12 pedigree. h The top boxplot was the transcription abundance of CRI12-like and Parent#2-like (with GhKHCP deletion) genotypes in Wang_2017 cohort (n = 195). The bottom boxplot was the disease index of 2 genotypes (Wang_2017 cohort; n = 195). P value was calculated by t-test and “***” represents for P ≤ 0.001. Boxplot center, median; bounds of box, lower quartile (Q1) and upper quartile (Q3); minima, Q1 − 1.5 (Q1–Q3); maxima, Q1 + 1.5 (Q1–Q3). i The conceptual illustration for bottleneck effect in cultivar-driven strategy. Orange segments represent for segments with low hereditary stability