Meta-transcriptomics using ContextMap

Recently, we presented ContextMap (Bonfert et al., 2012), a novel approach to mapping RNA-seq reads based on the concept of a so-called read context. A context is defined as a set of reads all originating from the same stretch of the genome, indicating that they all correspond to transcripts of the same gene. The original implementation we presented was focused on refining initial mappings provided by other mapping approaches such as, e.g. TopHat (Trapnell et al., 2009) or MapSplice (Wang et al., 2010), and improved the accuracy of the initial alignments considerably. In the following, we briefly describe the essential steps of the original ContextMap implementation. Details can be found in the original paper (Bonfert et al., 2012) also available at the Supplementary website (http://www.bio.ifi.lmu.de/ContextMap). With the initial mappings as input, ContextMap operates in three steps: