BAM file header error - ValueError: invalid coordinates:

Question

[rbronste]

Hi everyone,

Using msCENTIPEDE which is the python version for TF foot printing. The examples given here (https://github.com/rajanil/msCentipede) work perfectly however when I replace their BAM files with mine I get the following error while attempting to learn model parameters:

loading motifs ... num of motif sites = 471 loading read counts ...

Traceback (most recent call last):
  File "call_binding.py", line 344, in <module>
    main()
  File "call_binding.py", line 337, in main
    learn_model(options)
  File "call_binding.py", line 70, in learn_model
    for bam_handle in bam_handles])
  File "/Users/rbronste/msCentipede/load_data.py", line 119, in get_read_counts
    sam_iter = self._handle.fetch(reference=chrom, start=left, end=right)
  File "pysam/calignmentfile.pyx", line 896, in pysam.calignmentfile.AlignmentFile.fetch (pysam/calignmentfile.c:10127)
  File "pysam/calignmentfile.pyx", line 821, in pysam.calignmentfile.AlignmentFile.parse_region (pysam/calignmentfile.c:9672)
ValueError: invalid coordinates: start (131130842) > end (130694993)

I assume the difference is in the BAM files themselves, so maybe the headers need to be changed so the tool works. Any ideas would be really helpful. Thank you.

Rob

[jonasmst]

EDIT: Sorry, didn't read the stack trace. It's found a start position that is greater than a stop-position and can't make sense of it, by the looks of things. From the documentation: Start corresponds to the first base of the core motif for + strand motif instances and the last base of the core motif for - strand motif instances. In your BAM-files, is the start-positions on minus-strands greater than stop-position? If so, I'd try switching values for start- and stop-positions for - strand entries.