[Geuvadis_rna_analysis] GEM mappings

Dear all, we noticed some inconsistencies in the read lengths of the GEM mappings (I didn't check the BWA mappings, but it's probably the same). Some samples appear to have up to 76 bp matched, while other samples only have 75 bp. In regard to comparability this might cause some (probably very little) differences between the samples. i.e. NA20589.1.M_111124_3.bam has 75 bp NA20812.2.M_111216_6.bam has *76 bp* NA20760.3.M_120202_5.bam has 75 bp NA20783.4.M_120208_6.bam has 75 bp NA20768.5.M_120131_1.bam has 75 bp NA20798.6.M_120119_6.bam has 75 bp NA20803.7.M_120219_1.bam has 75 bp We checked some more files for institute 2, and they seem to have generally 1 bp more than the others. We're also a bit lost regarding GEM's quality score. We would like to filter reads for good mapping quality, we're just not sure how to accomplish this. Below are two read pairs as an example. The fist pair has a mapping quality of 180. The second read however was terribly aligned. The second pair has only a quality of 99, while the reads mapped much better (at least I would say so). We're not sure what's the reason for this. Curiously, both reads of a pair get the same mapping quality. This could be by intention (i.e. it's always the pair quality rather than the read quality) or, which could also explain the quality differences for the reads, both reads of the pair get the quality of the first mapped read. We're thankful for any suggestions on how to filter for 'good' mappings. best wishes, Matthias & Daniela -- Matthias Barann Institute of Clinical Molecular Biology Christian Albrechts University Kiel Schittenhelmstr. 12 D-24105 Kiel, Germany m.barann(a)ikmb.uni-kiel.de +49 - (0)431 - 597 8681 (office)