paperKB
coga / coga-kb
Help
Sign in

Chunk #29 — Results and discussion — Comparing bias across technologies — Bias on human samples

Source
Characterizing and measuring bias in sequence data.
Embedded
yes

Text

The human genome is far larger and more complex than the previously analyzed microbes and contains many examples of all of the motifs, as well as the 1,000 bad promoters (Table 1). We generated slightly more than one-fold coverage on the Ion Torrent PGM platform and 120-fold coverage on Illumina HiSeq. We also analyzed a 79-fold coverage data set generated by Complete Genomics. Complete Genomics sequencing, like Illumina and Ion Torrent, uses amplification in its process. We did not analyze the performance of Pacific Biosciences on human samples because, at the time of these experiments, the system's throughput made it impractical to generate sufficient coverage. To maximize comparability and avoid misinterpreting biological variation as sequencing variation, all data sets utilized the well-studied NA12878 sample [32] and were aligned to the Human Genome Assembly 19 (GRCh37) reference.