I have the data as shown below: (the samples are 200 app.)
SNP1 SNP2 SNP3
A/A C/G T/T
A/C G/G A/T
I have created the haplotypes using the Haploview (Haploview can provide me with the estimated value in percentages) But if, I want to know the exact number of each haplotype in the sample, how can I have that?
For example in haplotypes tab, the GCT (0.285) is written, so it means the frequency of GCT is 28.5% (as I have read somewhere - i just have to dig out that article). So how could be the actual number of haplotypes will be 0.285?
Firstly, without physical phasing, there is no way to know exactly the number of each possible haplotype. Statistical phasing only gives you probabilities, not integer numbers. Secondly, I think zx8754 is quite clear that number=frequency*totalNumber.
You can output haplotype frequencies as text file, then have a look at manual for explanation:
Haplotype Text Output File Haplotype output shows a block, its
markers, the haplotypes and their population frequencies, the
crossover percentages to the next block and the multiallelic D prime.
Crossover percentages are shown as a matrix with this block's
haplotypes as the rows and the next block'shaplotypes as the columns.
An example might look like:
In this example, the first block has 4 markers with 3 haplotypes
displayed and the second block has 3 markers and 3 haplotypes. The tag
SNPs for each block are (3,4) and (10,11) respectively. The crossover
percentage matrix can be read as follows: 80% of all samples have the
pattern 3312-441, 3.1% have the pattern 1144-441 and so forth.
Is it not frequency(percentages) multiplied by number of samples?
Thank you so much for your reply!
For example in haplotypes tab, the GCT (0.285) is written, so it means the frequency of GCT is 28.5% (as I have read somewhere - i just have to dig out that article). So how could be the actual number of haplotypes will be 0.285?
I hope I have explained it well :)
Firstly, without physical phasing, there is no way to know exactly the number of each possible haplotype. Statistical phasing only gives you probabilities, not integer numbers. Secondly, I think zx8754 is quite clear that number=frequency*totalNumber.