readAAMultipleAlignment does not correctly handle Stockholm-formatted MSAs generated by HMMer, generating the "alignment rows out of order" error.

On looking through the source, I think the problem arises from read.MultipleAlignment.splitRows() not correctly handling alignments where each row of the alignment can further be annotated by a comment line (e.g. "#=GR ...").  alnlines in the existing code (line 304 in MultipleALignment.R in github) expects blocks of all rows of the alignment to be on adjacent lines.  If they are interspersed with valid #=GR comments as in HMMER output, it incorrectly surmises the file consists of blocks of one row each and reports an error immediately.

A short term workaround is to strip out ^#=GR lines with sed or some similar tool before loading the sequence:-

Typical  example of alignment block that invokes error:-

SRR4381490.34595612_1_1/1-33         ---
#=GR SRR4381490.34595612_1_1/1-33 PP ...
SRR4381490.33050863_1_6/1-33         ---
#=GR SRR4381490.33050863_1_6/1-33 PP ...
SRR4381490.38714849_1_6/1-33         ---
#=GR SRR4381490.38714849_1_6/1-33 PP ...

Equivalent example which is accepted:-

SRR4381490.34595612_1_1/1-33         ---
SRR4381490.33050863_1_6/1-33         ---
SRR4381490.38714849_1_6/1-33         ---

Regards,

David Huen