In Silico Pcr - Small Coverage

Question

[Pawel Szczesny]

I have a set of primers designed by somebody else on the basis of Arb Silva SSU database (specifically, all are matching different fragments of 16S rRNA). The issue I have is that I cannot reproduce the depth of the coverage these primers should have according to the papers describing them with anything else than TestPrime server, Arb Silva's own software. I've tried e-PCR and MFEPrimer2.0, both with quite liberal conditions. I attempted to catch the sequences these primers _have_ matched in the real world PCR (I have amplicons already sequenced) and still I don't get enough coverage (like 5-10% of the things that are there). Is there anything I'm missing here?

[Istvan Albert]

Perhaps you could blast the primers against the known targets and see whether they are there at all in the first place, then check for correct orientation/pairing - just eliminating some possible problems. I am kind of thinking out loud.

[Pawel Szczesny]

Thanks for the suggestion. Blast (adjusted for short sequences) shows correct number of sequences matching the primers. I have tested all possible versions of primers (reversed, complement) I am sure I have the correct orientation and pairing.

Answer 1

[Pawel Szczesny]

The problem is solved, although not in a way I've hoped for. I've tried the third software, ipcress (part of exonerate package from Guy Slater at EBI; it's in the repositories of all major linux distribution). It gives correct results under liberal conditions (certain number of mismatches allowed). Still I have no idea why two other programs failed to give comparable results.