Figure 2.

The graphic model summarizing the parameters and data in our BM-Map algorithm. In the model, e and g are the observed data, representing the mismatch patterns of multiread and related unique reads, respectively; α is sequencing error rate by considering all the unique reads in the dataset; β is the probability of mismatch due to hidden nucleotide variations; q is the probability of mismatch based on α and β; and Z is the prior distribution of reads mapping, which is set in proportion to the numbers of unique reads associated with competing loci.

Yuan et al. BMC Genomics 2012 13(Suppl 8):S9   doi:10.1186/1471-2164-13-S8-S9