Especially with RNAseq data, one may want to output reads with V/J similarities,
but not random reads for other parts of the genome.
The new behavior of the '-u' option is to return these interesting reads.