Hi Sergei,

I am trying to identify sites under positive selection in a set of sequences that I simulated. The simulation was run so that all sites and all branches have an omega=1 (therefore, no sites should be detected as being under positive selection).

I ran this data set in PamL (model 0, nsites 2) and, as expected, no sites were found under positive selection.

In HyPhy, I am using dNdSRateAnalysis.bf followed by dNdSResultProcessor.bf as shown in the exercise on the InfluenzaA_H3_Random35 example. Using a BF threshold of 50 I find 827 codons (total number of codons is 1000) under positive selection. Am I doing something wrong or is there a reason for why HyPhy finds positive selection where there shouldn't be and where PamL doesn't find it?

Thanks,
fabia

A partial example of my final output:
Counter      Site Index      Log[BF{NS/S>1|s}]
1      1      5.38119
2      2      7.41881
3      3      7.12305
4      4      9.64254
5      7      6.73498

Hi Fabia

Can you post the alignment that gave the odd results?

Sergei

Mhmm, I am trying but it gives me an error. I sent it directly to your email.

Fabia