Short read data and variants detected per sample.

Short read data for the five samples was run in three lanes on the Illumina HiSeq 2000 platform resulting in roughly 200 million read pairs per lanelane (data available from sequence read archive associated with BioProject accession: PRJNA246137). During our filtration process it is possible some reads lose their mate while in cases where the reads overlap significantly, they are merged into one read. Therefore it is more useful to use the term fragments (the sum of pairs and single reads) as opposed to reads or pairs. In the range of 54–61% of fragments aligned successfully to our transcriptome reference sequences. The synthetic data set having been generated exclusively from the reference sequences aligned completely.†The synthetic reads were not filtered, this number is the total number of synthetic reads generated.