Fig. 2

Unsupervised grouping of 250 kb fragments derived from the pseudomolecule-anchored portion of the N. tabacum genome by polyCRACKER using repetitive k-mers and intact repeats. (a and b) PCA of repetitive k-mer count matrix of k-mers versus 250 kb genome segments for N. tabacum. S and T genomes are colored green and blue respectively, ambiguous segments are labeled red. a Sequences labeled by similarity to the known progenitor-like species. b Sequences labeled by polyCRACKER’s k-mer analysis. Note that a and b are nearly identical. c Spectral embedding of N. tabacum dimensionality reduced information, in which edges (grey lines) represent shared repetitive k-mer profiles) that connect sequences. Sequences labeled by similarity to the known progenitor-like species. d-f Analogous unsupervised grouping as above, but using informative and differential repeats rather than k-mers. Color labels as described above. d Sequences labeled by similarity to the known progenitor-like species. e Sequences labeled by polyCRACKER’s repeat analysis. Note that a, b, d and e are nearly identical. f Spectral embedding of N. tabacum dimensionality reduced information, in which edges (grey lines) represent shared repeat element profiles that connect sequences. Sequences labeled by similarity to the known progenitor-like species. Note the increased number of ambiguous subsequences in f as compared to c is due to the fact that there are fewer repeats than k-mers in any given scaffold. This is analogous to the effect of increasing k-mer size substantially