A probabilistic framework for aligning paired-end RNA-seq data
Yin Hu; Kai Wang; Xiaping He; Derek Y. Chiang; Jan F. Prins; Jinze Liu
Bioinformatics 2010 26: 1950-1957  

Recent News

08/18/2010: MapPER v0.12 is released. The bug in inferred CIGAR string in the output SAM file has been fixed.

07/26/2010: MapPER v0.11 is released. The inferred CIGAR string in the output SAM file may have 1bp shift (~1-2% junctions might suffer from this bug). We will address this in the next version. The junctions in junction_support folder are correct.

07/22/2010: The next verson of MapPER will address several format issues. Please stay tuned!

05/06/2010: MapPER v0.10 is released.




The MapPER method has been fully integrated within the MapSplice pipeline. For best results, we highly recommend using MapPER via MapSplice.

For paired-end read, the standard SAM format refers to pairing information by specifying at the read record of one end the mapped location of the other end read. This interpretation is hardly suitable when the data being addressed has multiple hits on both ends. In order to avoid listing all pairing combinations within sam files, MapPER takes an input file from MapSplice that has slight difference with the standard SAM format.

This difference ONLY exists at the input to MapPER provided by MapSplice. The output format of MapPER follows the standard SAM format as much as possible.




RNA sequencing using the paired-end protocol is a cost-efficient way to sample transcript fragments longer than the sequencing capability by sequencing only the ends. This software implemented a probabilistic framework to predict the alignment of each transcript fragment to a reference genome for each PER.  The alignment chosen is determined by maximizing the likelihood of all PER alignments through an expectation maximization method.