Fig. 7.
Considering only long genes (>300 aa) does not change the result that between-genome variance accounts for most of the variance in the first two dimensions. (a) Same as Fig. 1a. (b) The corresponding graph of singular values made for a data set containing only genes longer than 300 amino acids, where the sampling was 200 genes per genome. The difference in absolute magnitude is due to the use of different data set sizes (400 genes per genome for all genes in the graph in a). (c) Same as Fig. 1b. (d) The corresponding graph of the contribution of between-genome variance to overall variance in each dimension for a data set containing only genes longer than 300 amino acids, where the sampling was 200 genes per genome. Between-genome variance in the first two dimensions contributes even more to overall variance in those dimensions when only long genes are considered.