Overcoming PCR Primer Bias with MiSeq?

[emyoung90]

Hi, I am a grad student new to MiSeq and HTS in general. I'm trying to figure out which sequencing kit/method to use for my project, and I'm getting some mixed responses. I have spoken to 5-6 people on the matter (some of whom work at Illumina), and both TruSeq Custom Amplicon sequencing and Nextera XT have been suggested - I'm just trying to figure out the best way to go about this, since my budget is limited.

Background: I am trying to detect co-infections of malaria strains in host DNA samples. There are 5 lineages/strains that I am looking for, all of which we have unique sequence fragments for from PCR/Sanger methods.

Issue: The primers and PCR/Sanger sequencing protocol we have are great for amplifying and identifying the dominant lineage present, but the issue is that the primers are biased - they will amplify some lineages over others for various reasons. We want to study co-infections of multiple lineages in a single sample, but our PCR/Sanger methods cannot determine co-infection due to that bias.

The sequences we have are all from the malaria cyt-b gene (mtDNA), fragments are about 500-550 bp long, and the sequences are fairly conserved between lineages - they average about 10 nucleotides different from each other. As such, I can't develop lineage-specific primers for PCR, as the sequences are too similar.

Goal: To use a high-throughput technique to target that same region (so that we can use our lab database to identify the lineages). I had thought to use a
generic malaria probe (one that will bind with any malaria present, but not with any host DNA), but I need to ensure somehow that the same primer bias described before won't negatively affect results. I would think that would be remedied by read depth, but I admit that I am new to HTS-related techniques so I'm not sure.

I had been considering TruSeq Custom Amplicon sequencing, but because my target area is so short (~600 bp), that method would probably not be cost effective. Can anyone give any advice on this? Thanks.

[thermophile]

You could do direct amplicon sequencing, the way that most people are doing 16s for bacterial communities. (see Kozich 2013 or Caporasso 2012). Depending on how many samples you have, you may want to design your own primers or use a 2-step method (Lange 2014)

I run a core facility that specialized in amplicon sequencing, you can check out my website for some documentation and visualizations. mars.uconn.edu

[fanli]

I would tend to agree with @thermophile that custom primers would be the way to go here. If you are on a limited budget, Illumina kits tend to be quite expensive on a per-sample basis.

If you are trying to detect co-infections, then yes, simply having tens or hundreds of thousands of reads per sample of any malarial DNA should in theory be sufficient. If you are trying to measure co-infections in a quantitative sense, then I would be more worried about primer bias. Then you could do something like a synthetic control of known lineage amounts, and see how reproducibly you can recover this ground truth using your assay.