Distribution of TE copies according to their dinucleotide mutation bias. Mutation rates were calculated using RIPCAL by comparing each TE copy with a Ti/Tv > 2 with the TE consensus sequence or the full-length genomic copy with the highest GC content through multiple-alignments. Y-axis: percentage relative to the total number of copies used in RIPCAL analysis. - Colored bars correspond to the percentage of copies with expected RIP* and dinucleotide preferentially used (>1/3) in CN- > TN and (cNG - > cNA) mutations. - Black bar: percentage of copies without expected RIP* (Ti/Tv > 2). - Gray bar: percentage of copies with expected RIP* but no evidence of dinucleotide bias. * RIP expected: Ti/Tv > 2. Note: more than one dinucleotide bias could be observed in a single TE copy.