A binning tool to reconstruct viral haplotypes from assembled contigs

Jiao Chen; Jiayu Shang; Jianrong Wang; Yanni Sun

When each position of the underlying haplotypes can be covered by at least one contig, d is equal to or larger than the number of haplotypes. Note that the common regions between different haplotypes are regarded as different positions and thus should be covered by different contigs. This conclusion can be proved by contradiction easily. Fig. 2. (b) shows the contigs satisfying the conditions in the ideal case. There are three haplotypes with different abundances. They only contain mutations at three positions that are far away from each other. Because of the long common regions among them, assembly programs usually won't be able to recover all the three genomes. Instead, they can generate short but correct contigs. In Fig. 2. (b), each position in the three haplotypes is covered by at least one contig. In this case, all the windows have depth of at least 3. If every position of a haplotype is only covered by one contig, the windows will have depth N . As some positions can be covered by multiple contigs, the overlaps between contigs contribute to window depth larger than N . For example, in Fig. 2. (b), the middle window contains the overlaps between two contigs from each haplotype and thus has depth 6. In this ideal case, we can choose the smallest window depth as the number of haplotypes in a sample.