Extended Data Fig. 5: Very long (>750 bp), near perfect, inverted repeats can lead to false positives.

(A) Alignment of inverton NZ_CP025371.1:2124719-2124870-2125316-2125467, with its invertible sequence flipped, against the B. pertussis genome leads to perfect alignment of flanking and IR regions as expected. ‘Reference genome’ refers to the B. pertussis reference genome sequence. ‘Inverton reversed’ refers to the putative inverton sequence and flanking sequence, with the invertible sequence inverted. Red dashed lines indicate boundaries of the invertible sequence, black dashed lines indicate boundaries of the inverted repeats as detected by einverted, and purple dashed lines indicate the true boundary of inverted repeats. (B) Alignment of the reverse complement of the entire inverton NZ_CP025371.1:2124719-2124870-2125316-2125467 with its invertible sequence inverted and flanking sequence, against the B. pertussis genome leads to near perfect alignment (6 mismatches) spanning far into the flanking sequence to the true boundary of the inverted repeats, allowing for reads to map regardless of inverton orientation. (C) Example with toy nucleotide sequences. Red nucleotides indicate mismatches.