From my very limited understanding, the age of a polymorphism (and perhaps some mutations that do not compromise fertility) may be inferred from allele frequency, as suggested from the 1k Genome Project (Nature 491, 56, 2012).

Dear Albino, I detected this mutation in 1 family. This mutation was absent in 100 healthy volunteers. Could your approach still be considered or should the age of the mutation be determined differently. However, I am not familiar with inferring allele frequency. Would it be possible that you could help/assist?

I would probably proceed as follows, but please remember my expertise is limited in this field.

1) Check if the mutation has already been described or at least reported in the HGMD database (http://www.biobase-international.com/product/hgmd). You may need an institutional subscription to access the site.

2) Check that the mutation has not been listed in the 1000 Genomes Project, which lists >30 million apparently benign polymorphisms.

3) Assuming that your analysis is representative of the population, the allele frequency that you currently have is expected to be less than 1%. According to the Nature paper, allele frequencies of ~0.5% appear to reflect recent polymorphisms. But I do not know how your phenotype would complicate the interpretation. In the meantime, I'll think a little more and hope someone else has better suggestions. Best.

Dear Albino, I checked all the available databases for the presence of this mutation prior to calling it novel. As of the prevalence estimate. We used the same population. The observed mutation seems to be strongly associated with the clinical phenotype, eg bleeding tendency.

Looking forward to any complementary suggestions/insights.

Thanks, René

I am no expert on this, but I don't think this is possible. You have a single family with a private mutation. From what I understand, people use the haplotype around the mutation and compare this haplotype among different mutation carriers. By estimating mutation frequencies in e.g. non-coding regions around the mutation, you can probably calculate the age of a mutation. How far the mutation has spread in a population also helps calculating the age. Check out papers on lactose intolerance / lactase persistence. I believe that's how they came up with the 10000 year age of the mutation.

You cannot compare the mutation haplotype to any other family and the mutation has not spread (does it influence reproductive fitness?), so I would not know how you could do it.

Especially if the mutation does not impact reproductive fitness much, a private mutation in a single family should be very recent.

A mutation that causes a phenotype subject to selection, such as any bleeding disorder would likely to, cannot have the age of the mutation determined by allele frequency alone.  The strength of selection will have a strong impact on allele frequency.  If the mutation has never been detected before in any other cohort, then it is possible that it is unique to this family.  In that case, your best bet is a really extensive, well-researched family pedigree.  If the mutation has been detected in others, try to find a common ancestor between the groups or do an estimate of common ancestry based on an linkage disequilibrium map of the region (this can also be useful if the SNP only occurs in this family).  Best of luck.