Currently the program searches for the exact string DS:Z: in the whole line from stdin to restore the information using added optional fields from previous runs of the program. This creates problems when the pattern is present in fields other than optional fields. We encountered this when trying to process a bam file that had one single such problematic read in the whole ~60GB bam. To reproduce the issue, the read is pasted below:
ERR566093.277697703 0 2 158103001 37 55M * 0 0 AGCACTCATGTTGTTCTGCGTGACAACTTTGCTTAGGCCGAGCCCTACAGAAACT ]R]TUNR3Z]I]]S];S]6R[AHP??UWTS]Z]]DS:Z:Q5YKUPT[]]]LOP[M X0:i:1 X1:i:0 MD:Z:55 RG:Z:fooXG:i:0 NM:i:0 XM:i:0 XO:i:0 XP:i:1 XT:A:U
outputs:
Traceback (most recent call last):
File "pmdtools.0.60.py", line 349, in <module>
PMDS= float(line.split('DS:Z:')[1].rstrip('\n').split()[0])
ValueError: could not convert string to float: Q5YKUPT[]]]LOP[M
when the program is run only with the --deamination option.
Currently the program searches for the exact string
DS:Z:in the whole line from stdin to restore the information using added optional fields from previous runs of the program. This creates problems when the pattern is present in fields other than optional fields. We encountered this when trying to process a bam file that had one single such problematic read in the whole ~60GB bam. To reproduce the issue, the read is pasted below:outputs:
when the program is run only with the
--deaminationoption.