duplicate records for inversions

[sb43]

I recently ran BRASS for a sample and found something I think is unusual in the final calls. Three inversions with scores for reconstruction had duplicated lines in the final output BEDPE, e.g. the line starting “chr 6 100392225”. Usually, these sorts of things represent the reciprocal breakpoint called in the other direction. When I examined the output though, the strandedness was identical on both lines (see output below). This seemed to be a problem when the breakpoints were identical in both directions.

Is this a bug or an expected behaviour that I have simply misunderstood?

chr6 100392225 100392227 chr6 105302597 105302599 248 6 - + SAMPLE1 249

chr6 100392225 100392227 chr6 105302597 105302599 249 4 - + SAMPLE1 248

chr7 11148335 11148336 chr7 41835781 41835782 250 5 + - SAMPLE1 251

chr7 11148335 11148336 chr7 41835781 41835782 251 8 + - SAMPLE1 250