We have previously interpreted human-mouse xenograft transplant16 and co-culture20, 33 data using rational methods, but we wished to perform a more uniform comparison of several potential methods. Therefore, we chose public datasets from cell types appropriate for our studies as a consistent source. To generate ‘hypothetical’ RNAseq data with pooled reads from both human and mouse, we randomly selected a total of 9,500,000 paired-ended reads from two FASTQ files with different percentages of human content (0, 10, 50, 90 or 100%). To track the source of each read, a prefix of ‘human-’ or ‘mouse-’ was added to the read ID in the FASTQ records.
