A Hidden Markov Model to estimate population mixture and allelic copy-numbers in cancers using Affymetrix SNP arrays

Table 4 Compararison of two estimation methods on 40 simulated samples

		Accuracy in %		Average posterior probability
Method	Number of samples	Average	stdev	True state	False state
One-array
- only normal heterozygous calls	20	99.495	0.058	0.997	0.812
- all calls	20	98.548	0.147	0.992	0.845
All-array
- only normal heterozygous calls	20	99.552	0.042	0.997	0.806
- all calls	20	98.653	0.150	0.992	0.830

Here the one-array method (transition parameters are estimated for each sample and are the same for all chromosomes) and the all-array method (transition parameters are the same for each chromosome across all samples) were compared. For each method, the accuracy of the HMM on samples where all SNPs are heterozygous in the normal sample and the accuracy of the HMM on samples where the call distribution is the same as a HapMap sample (all calls) were also compared. Moreover, an average of the posterior probability for all SNPs correctly recovered or not is given.

ISSN: 1471-2105