Aligning virus reads against reference genome

[flacchy]

Hi all,

I manage to select some reads from my metagenome illumina analyses. From my blastn analyses I am sure they are from a specific virus but when I try to map those against the reference genome neither tophat2 or bowtie2 allow me to have an alignment... I think is because of big variation between the reads. So... my question is ... is there any other program or way I can map my reads against a reference genome? The goal is to identify what part of the virus is present... so other times I just used Illumina and then samtools and IGV but this time I might need something different and I can't figure out what program will suits better.
If anyone could help it will be great.

Thanks

F.

[GenoMax]

Have you tried to blat the reads against the viral genome?

[flacchy]

No I haven't but I will look at the manual even if it doesn't seem an alignment tool... so in total now I have 300 reads of a specific organism and I want to map it against a reference to see exactly where they align. Do you know if is possible?

[GenoMax]

No I haven't but I will look at the manual even if it doesn't seem an alignment tool... so in total now I have 300 reads of a specific organism and I want to map it against a reference to see exactly where they align. Do you know if is possible?

Blat is an alignment tool much like BLAST but fast. It should be easy to map 300 reads to the viral genome and best of all you do not need to create indexes etc. Just two fasta files (one with your reads and the other with the viral genome) is all that is needed for Blat.

[rhinoceros]

Why don't you just blast the reads against the reference genome? Blast is more accurate than blat and 300 reads is nothing.. should take a few minutes top.

[westerman]

@rhinoceros: flacchy already used blast as per the second sentence "... From my blastn analyses ..."

Bowtie2 is created to map reads quickly. If you are not getting any alignments, or only non-significant alignments, when using the --very-sensitive-local (or any of the local settings) then you simply do not have reads that are from your virus no matter what blastn indicated.

blat is a useful tool and sometimes more sensitive than blast for some types of searches. Also running tblastx instead of blastn can, in some circumstances, provide a more sensitive search.

But really, I am suspecting that you either don't have the proper reference or your reads just are not from the virus.

[rhinoceros]

@rhinoceros: flacchy already used blast as per the second sentence "... From my blastn analyses ..."

Ah yeah, true. But then, hasn't he already achieved his goal (identify what part of the virus is present)?

[westerman]

@rhinocerous: Yes flacchy should be able to parse the loci from blastn. Which is why I think he is somehow barking up the wrong alley (for the non-American's, this is an idiom for "going in a wrong or mis-guided direction"). Without more details of why he thinks that reads belong to a given virus and, as you say, why he can't get his information from the blastn results ... well ... it will be hard to give more help.

[chriswong]

If you are using Illumina, you can try the new PathGEN PathSEQ tool that has been made available on BaseSpace. https://basespace.illumina.com/apps/...PathSEQ-Virome
This tool will be able to determine if there are any viral sequences in the dataset.