Closed kingralph80 closed 4 years ago
Hi, I added a new option (-unique) to only output unique alignments (version 1.0.21). It should remove all redundant alignments. And I've changed ref and qry to ref. and qry.
I fixed a bug in determining the uniqueness of alignments. Please update GSAlign to version 1.0.21. Thank you! Non-unique alignments in MAF file will be annotated with "a score=1." You could skip those alignments in variant calling.
Hi. Thanks. How does -unique work? So it will only put out the best alignment? What if two alignments get the same score?
It simply checks if a query fragment have multiple alignments against the reference sequence. Just like short read alignments, some reads may have multiple hits if they come from a repetitive sequence region. So given two alignments, say Aln1(qry_pos1, qry_pos2, ref_pos1, ref_pos2) and Aln2(qry_pos3, qry_pos4, ref_pos3, ref_pos4), if qry_pos1 = qry_pos3 and qry_pos2 = qry_pos4, then the two alignments are considered not unique. In such cases, Aln2 will be removed and the alignment score of Aln1 will be assigned to 1 if -unique is set. However, if two alignments are highly overlapped, that is they share a large portion of query block or reference block, they are also considered not unique. In such cases, the shorter one will be removed.
If two alignments get the same score, it is not necessarily they are not unique. Uniqueness are based on the alignment regions, rather than the alignment scores.
Thank you!
HI,
thanks for updating GSAlign so quickly. Its really an amazingly fast Genome Alignment tool!
When we checked the VCF output from GSAlign, many entries are duplicated. In addition, we found many additional SNPs in alignment that did not have any SNPs. It is coming from multi-alignments?
In the past you added ref and qry to allow easy handling of same chromosome names. This was a great addition, however, most tools such as mafToPsl or mafToAxt only support . instead of .Could you change the ref and qry_ to ref. and qry. ?