quaidmorris
commented
2 years ago Population frequency of a mutation is the percentage (or proportion) of cells in the sample that have the mutation. So if p% of cells have the heterozygous mutation then the VAF, when expressed as a percentage will be p/2% because each cell with the mutation contributes p% of the alleles to the total number of alleles from that locus in the population, and 1/2 of the alleles that these cells contribute have the mutation. This is assuming that the loci in both the normal and mutated are diploid.
If the VAF (as a percentage) is p/2% and there are di reads mapping to that locus, then, on average, we would expect di x p/2 of them to contain the mutation.
Cancer cell fraction (CCF) normally means the proportion of cells in the population that are cancerous. So if the mutation in question is a clonal mutation (I.e., is present in all cancer cells) then the CCF will be p% and the VAF of that clonal mutation will indeed be p/2%.
However, the relationship between population frequency and VAF still holds if the mutation is subclonal, I.e., p% < CCF
Q
On Thu, Jul 7, 2022 at 10:33 AM HushWay @.***> wrote:
Dear sir,
Your article said that "we create an equivalent pseudo-SSM with population frequency φi by adding an SSM to the dataset with total reads di and variant reads di × φi/2 rounded to the nearest whole number", but this really confused me because only cancer cell fractions rather than population frequency of heterozygous mutation are twice of VAF. If I did not make a mistake, your population frequency refers to cell fraction of all cells including normal cells. How to explain the formula?
Thanks for your attention, Liuwei
— Reply to this email directly, view it on GitHub https://github.com/morrislab/phylowgs/issues/137, or unsubscribe https://github.com/notifications/unsubscribe-auth/ABIZI5SZJKPVRRT5VYKI3NLVS3TE3ANCNFSM525UVUWA . You are receiving this because you are subscribed to this thread.Message ID: @.***>
--
Quaid Morris, PhD Member, Computational and Systems Biology, Memorial Sloan Kettering Cancer Center CCAI Chair, Vector Institute Faculty (on leave)
HushWay
Dear sir,
Your article said that "we create an equivalent pseudo-SSM with population frequency φi by adding an SSM to the dataset with total reads di and variant reads di × φi/2 rounded to the nearest whole number", but this really confused me because only cancer cell fractions rather than population frequency of heterozygous mutation are twice of VAF. If I did not make a mistake, your population frequency refers to cell fraction of all cells including normal cells. How to explain the formula?
Thanks for your attention, Liuwei