Analysis of viral amplicons with hypervariable regions

[Himalaya]

Analysis of the reads by codon would be a very helpful step forward. I'd be interested to try the tool you are developing once it is released. I was surprised that AVA did not have such a capability

Have you trialed any other open source software packages?

Cool, Will let you know about the tool. Sorry for late reply. More update on the tool. Now the tool is also able to remove the sequencing errors or pcr error based on the codon space alignment and output the error free reads in separate file. It also generates unique sequences (not consensus sequence) applying the filter such that the unique sequence must have 3 forward sequences and 3 reverse sequences all considered in codon space. In this way the unique sequence is very much error free.

I did try SHORAH and RC454.

[Mr Jefferson]

Analysis of the reads by codon would be a very helpful step forward. I'd be interested to try the tool you are developing once it is released. I was surprised that AVA did not have such a capability

Have you trialed any other open source software packages?

[Himalaya]

Hi Mr Jefferson
I am researching in HIV Drug resistance mutations (DRM) in HIV quasispecies from infected patient to develop a tool for quick DRM detection. Initially I also tried to use AVA but it's difficult to configure out how AVA makes consensus sequences after reference mapping. We totally dropped out using AVA now and my supervisor team is developing a new tool for mapping on codon space. Usually mapping is done in nucleotide base and indels in sequence reads specially at homopolymer site or hypervariable site is difficult to identify. Mapping at codon space can identify the indels or sequencing error specifically.

[Mr Jefferson]

Hello all,

I have done some pyrosequencing on HCV samples with the amplicon spanning the HVR1. I also have a sample that contains a clonal sequence as a control for sequencing/homopolymer error.

From my limited background in the area it appears that the analysis software provided by Roche has limited capabilities when it comes to processing virus sequences. In particular AVA views the mutations as random which, in the context of a quasispecies population with a hypervariable region, isn't the case. Consequently it is difficult to define a consensus sequence that will capture the optimum amount of information from the data file.

Do any users on this forum have experience of handling 454 sequence data from virus amplicons? I note a recent paper (http://www.biomedcentral.com/1471-2105/13/S10/S6/) where the group developed algorithms to assist in the calling of true variants and that it an improvement on SHORAH. As yet though I haven't been able to implement it myself.

I be interested to learn of more examples where groups have developed their own solutions to the problem.