A major challenging problem using NGS sequencing data is the annotation of reads aligned at multiple locations. Most of the available frameworks resolve this situation by discarding these reads or by providing random annotations. Here, we propose to keep all the reads aligned to the genome, and to weight them by the number of mapping sites. Suppose a read can be aligned 5 times to the genome, for each mapping site, the read would be counted as 0.2, i.e. 1/5.
Chongjian Chen
2012-01-26
