Figure 2
From: A context-based approach to identify the most likely mapping for RNA-seq experiments
Definition of contexts and extension of alignments. A) A context is defined as a set of reads with overlapping mapping locations such that the minimum distance between any pair of reads from different contexts is d min . B) The initial alignments used as input to ContextMap are extended by re-aligning unmapped or multi-mapped reads to each context using a forward alignment from the read start towards its end and a backward alignment from the read end towards its start. Forward and backward alignments are combined to a full or split read alignment if the maximum mismatch criterion is met. (C) All alignments obtained at this step provide a list of potential splice sites. Additional alignments respecting the maximum mismatch criterion are then created for full and partial read alignments based on these potential splice sites. Accordingly, for any read aligning up to the first position s1 of a splice site it is investigated if the rest of the read can be aligned starting at the second position s2 of the splice such that only a maximum number of mismatches are created. These additional alignments are then added to the set of candidate alignments.