GATK -rf BadCigar (Reads starting with deletions)

[JezSupreme]

Please excuse any naive or ignorant questions here, but I'm still very new to bioinformatics. I'm trying to run some illumina sequence data through the GATK pipeline and have been running into problems with reads beginning with deletions. I've been running some initial commands on my .bam files to make sure GATK is OK with the files and while I can run CountReads successfully, I get a BadCigar error return when I run the CountLoci command. Of the 3 data files I'm looking at, all three return BadCigar errors, and all 3 errors are related to reads starting with deletions.

While I can run these files successfully by using the '-rf BadCigar' flag, this use raises a couple of questions:

1.) Why does GATK consider these reads to be in error when we have yet to map the InDels and conduct a realignment?

2.) The process of using the BadCigar flag seems to ignore the entire read, which is not ideal for my purposes. Is it possible to force the analysis to run and simply permit the bad cigars rather than ignore the entire read?

3.)The GATK doesn't seem to report which reads have been ignored (beyond the initial run of the file not using the flag, in which it error's out at the first bad cigar). Is it possible to have the GATK report each bad cigar as it is ignored in the analysis?

Thanks to anyone who can shed some light on these questions.