PAML NsSites sequence length influences significance?

Question

[susannak]

I ran an analysis on two different genes using NsSites in PAML. The first gene is 9 codons (27 bp) and the other is 389 codons (1167 bp). NsSites identified a handful of sites under positive selection for the longer gene and none for the short gene. I am wondering how much sequence length influences significance values?

is my short gene truely not under positive selection, or perhaps PAML isnt the best method for such a short sequence. Does anyone have any insight?

[pr]

If you found any answer concerning this issue, I'd love to hear it!