queryBases filter with pattern returns no matching reads

[adomingues]

I am not sure if this is a bug or if I am setting up the command incorrectly, but when trying to filter alignments based on the first nucleotide in the query sequence, no match is found. This what I tried (bamtools 2.5.1):

1. Using my actual query, where I am trying to get sequences that start with an "A"

bamtools filter -in WT_TAP.bam  -queryBases "^A*" | bamtools count
0

This is unexpected because there are plenty of sequences that should match this pattern:

samtools view WT_TAP.bam | cut -f 10 | egrep "^A" | wc -l
712681

2. testing a (full) sequence that exists in the file

bamtools filter -in WT_TAP.bam  -queryBases "GTTTGACGGCCTTGATGTGCACT" | bamtools count
5

which returns the expected output.

3. testing a subset of the above sequence with and without regex

bamtools filter -in WT_TAP.bam  -queryBases "GTTT" | bamtools count
0

or with pattern

bamtools filter -in WT_TAP.bam  -queryBases "GTTT*" | bamtools count
30667

Which returns the expected result. However, if I restrict it to only a single nucleotide, it again fails to return any match:

bamtools filter -in WT_TAP.bam  -queryBases "G*" | bamtools count
0

---

This is a bit puzzling. Could you please help me construct a query to select alignments whose query sequence starts with a particular nucleotide? Along the same lines, and related to this question, what is the best way to take in account the reverse-complemented sequences (0x10 flag)?