Fig 1.
An example of SNP matrices X and H relevant to the resulting haplotypes.
The red measures in X indicate sequencing errors. Each row of H demonstrates a specified haplotype sequence.
Fig 2.
The workflow of proposed method.
Fig 3.
An example of constructing and partitioning the hypergraph.
Si corresponds with the ith SNP, and the curves demonstrate the hyperedges. C1 and C2 denote the clusters which are obtained by hypergraph partitioning.
Fig 4.
Partitioning Ht of a three ploid genome.
The yellow parts indicate and the green parts demonstrate
. It must be pointed out that
.
Fig 5.
An example of updating the current haplotype based on the partitioning of the hypergraph.
Fig 6.
The algorithm of improving Ht.
Fig 7.
Two combinations of six possible combinations of the CutSet in three ploid form.
Table 1.
Average of reconstruction rate for haplotypes with length 100.
Table 2.
Average of reconstruction rate for haplotypes with length 350.
Table 3.
Average of reconstruction rate for haplotypes with length 700.
Table 4.
The effect of refinement phase for haplotypes with length 700 in diploid case.
Table 5.
The reconstruction rate and running time for the proposed method, H-pop, SCGD, HG, ARO and FCM applied to the experimental dataset NA12878 dataset provided by 1000 genome project.
Table 6.
Average of reconstruction rate for haplotypes with length 100.
Table 7.
Average of reconstruction rate for haplotypes with length 350.
Table 8.
Average of reconstruction rate for haplotypes with length 700.
Table 9.
The effect of refinement phase for haplotypes with length 700 in polyploid case.
Fig 8.
Comparison of reconstruction rate of the methods over high coverage data a) Diploid b) Polyploid.