I'm using VarEff to calculate the Effective population size using the function NatSizeDist. 'It calculates the estimates of effective size (Ne)
at a number of times from 0 to a certain time ago (given by the user), plots and
saves results on files.' According to what I've read it seems that one should use the harmonic mean to determine the Ne, but my harmonic mean does not change between runs using different mutation rates (I'm using microsats). The mean, however, does fluctuate. Could someone perhaps tell me why this could be the case and could one use the mean instead ( the mean values make biological sense)?
Any help would be appreciated!
Harmonic mean is often used to average speed or time, when data are heavily influenced by small values. But when a variable value is close to zero or zero, then harmonic mean can not be calculated. Harmonic mean also gives large values less weight, while gives small values greater weight than that given for the arithmetic mean.
Check if there is any zero value on your Ne data affecting harmonic mean calculations. Biologically seems for me that Ne zero values shouldn't have sense.
Hi Harold, thanks for the response and insight! The input file has zero's in it as the manual says that you need to specify all potential alleles between those of minimum and maximum length, so if you have a locus with 2 types of alleles and their lengths are 10 (frequency = 24) and 12 (frequency = 6) (number of repeat motifs), you have to mention the unobserved allele with 11 motifs. So your input file will say;
3 (alleles)
24 0 6 (frequencies at 10, 11, 12)
Do you have any experience with VarEff or know of any way to circumvent the issue?
- Badges
- Science method
- Similar topics
- Biological Science
- Ecology