Why do we remove primer sequences?

[greggiceton]

Hi

It seems to be the default approach that we remove PCR primers from our final sequence data. But given that no primer is 100% universal and with a low annealing temperature we might capture greater diversity, I am wondering why the standard approach is to remove this primer sequence and thus potentially lose some sequence diversity?

[sarvidsson]

You are talking about 16S amplicon sequencing? The actual primer sequence is not so informative as you force the amplicons to match it as you cycle on (most of the final fragments will carry primer sequence) - i.e. the sequence diversity caught by mismatch-hybridization due to low annealing temperature will be ironed out by the very primer sequence. Therefore you cannot assume the final sequence in this region to correspond to the actual diversity in the sample.

[greggiceton]

There are times when the penny drops so loudly it can be heard for miles around. This is one of those times. Thanks for the clear explanation.

[ATϟGC]

For both sanger and high-throughput sequencing(HTS), you should almost always delete primer sequences because they are synthetic. That is to say that they may, or may not, reflect the organisms sequence where annealing occurred. I can not think of any instances where I would not trim off primer sequences prior to an analysis (unless I was curios whether there is differential PCR amplification among primer pairs in a HTS experiment). I have seen some genbank accessions of sanger data where the primer sequences were not removed from the locus but most consider this to be a poor practice.