I am trying to gain confidence on my results. I have a graph contig that looks like this:
SP5 - SP34 -> SP45 -> FL53
In total there are about 12 short reads (100 bp long) that are sent to Velvet which outputs something. The problem is that I expected exact match (or at least next to exact) between the features and the assembled CRISPR. My particular spacer group only has one direct read, and the direct read is matched exactly, together with SP5, I am yet to test a full multiple aligner but you can see for yourself by playing with this script that even in smaller subreads there is little match.
fa = """ATTGTGGGGCAACTGAAACCCGACGAAGGCTGGCACGTTGCCCGCACCCGTATTGGCCCC
CTTTAAATTGTGGGGCAACTGAAACAACGTAGCAGCCTCGGCTGGTTCGTCAGGGTATTG
GCCCCCTTTAAATTGTGGGGCAACTGAAACTATGTATACGATGATCCATACACCGTAT"""
fa = fa.replace("\n","")
features = {
"DR": "GTATTGGCCCCCTTTAAATTGTGGGGCAACTGAAACTA",
"SP34": "TGTATACGATGATCCATACACCGTATT",
"SP5": "CGTAGCAGCCTCGGCTGGTTCGTCAGG",
"SP45": "CATGACGCCCTGATGGCGGAATACGCCCC",
"FL53":"CCAGATTAAACTTTGGCTCTCTATTACTACCAACTCAAAGCGACTACTTTAGATGATCTATGT"
}
print(fa)
for f in features:
print(f, fa.find(features[f]))
My questions are:
- Should I expect for the spacers to not be matched exactly, or is there an error to be suspected?
- Should I expect to not see the small contig subgraph structure reflected in the assembly?
- Is there a way to asses how far is my assembly from the predicted spacer graph?
- Bellow is the XML part describing my contig, can I expect that one single spacer is actually assembled connected to my only direct read (there is only one DR for this spacer graph).?
<contig cid="C1">
<cspacer spid="SP5">
<fspacers>
<fs drconf="0" drid="DR1" spid="SP34"/>
</fspacers>
</cspacer>
<cspacer spid="SP34">
<bspacers>
<bs drconf="0" drid="DR1" spid="SP5"/>
</bspacers>
<fspacers>
<fs drconf="0" drid="DR1" spid="SP45"/>
</fspacers>
</cspacer>
<cspacer spid="SP45">
<bspacers>
<bs drconf="0" drid="DR1" spid="SP34"/>
</bspacers>
<fflankers>
<ff directjoin="0" drconf="0" flid="SP53"/>
</fflankers>
</cspacer>
<cspacer spid="FL53">
<bspacers>
<bs drconf="0" drid="DR1" spid="FL45"/>
</bspacers>
</cspacer>
</contig>
Thank you!
I am trying to gain confidence on my results. I have a graph contig that looks like this:
SP5 - SP34 -> SP45 -> FL53
In total there are about 12 short reads (100 bp long) that are sent to Velvet which outputs something. The problem is that I expected exact match (or at least next to exact) between the features and the assembled CRISPR. My particular spacer group only has one direct read, and the direct read is matched exactly, together with SP5, I am yet to test a full multiple aligner but you can see for yourself by playing with this script that even in smaller subreads there is little match.
My questions are:
Thank you!