bwa alignement outside of the reference sequence

[cklopp]

Hi,

Using bwa (Version: 0.5.9-r16) we encountered cases where the proposed alignment position was outside of the reference sequence.
In the example hereunder the alternative alignment position of the read starts, on the forward strand, at position 145240300 on a chromosome of length 145240301

@SQ SN:CHR13 LN:145240301
@SQ SN:CHR14 LN:148515138
@SQ SN:CHR1 LN:295534705
@PG ID:bwa PN:bwa VN:0.5.9-r16
[bwa_aln_core] convert to sequence coordinate... 9.64 sec
[bwa_aln_core] refine gapped alignments... 4.18 sec
[bwa_aln_core] print alignments... HWI-ST314_0106:4:1202:17139:18375#0 16 CHR15 132557769 0 101M * 0 0 ACAGAATTCTATATCCAGTGAGAATATCCTTCAGGTATGAAAGTGAAAAGAGATTTTCTCAGATGAAGGAAAACCGAGTGAATTTGTTGCCAGTGGAGCTA JPJJS`QKQQKQKK[[QJRJJRJK`J`KbK[`Q`RJIIPPP^ceSI^I^IOOHIIOOOOacgINOa_aBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBB XT:A:R NM:i:4 X0:i:2 X1:i:1 XM:i:4 XO:i:0 XG:i:0 MD:Z:8G26A50C10T3 XA:Z:CHR13,+145240300,101M,4;CHR14,-18932,101M,5;

Any idea of what happens?

Cheers

Christophe

[swbarnes2]

bwa concatenates the reference sequences together, so if your read hangs off the end of its reference, that'll happen. But bwa should give it a 4 in the binary flag, so that's how you know the mapping isn't quite right.