Biomedical Engineering Reference
In-Depth Information
Fig. 8.1 Whole exome sequencing workfl ow. The DNA is fragmented, library is prepared, and
reads are generated by NGS instrument (i.e., HiSeq2500). Determining nucleotide calls (A,C,G,T
or N) along with error probabilities (Q score) is performed via a proprietary base calling algorithm
during the sequencing run . The FASTQ fi le is the raw data which contains the base calls and qual-
ity score per base. The major step in the analysis is the process of aligning the reads to the human
genome, which often takes few hours to complete. To assess the quality of exome sequencing QC
parameters, which include the number of reads aligned successfully and the depth of coverage
percentage on specifi c target region, are checked. The alignment procedure yields multi-sequence
alignments in BAM format. Next, variants are determined by comparison to an indexed reference
sequence. Statistical scores are computed to reduce false positive errors due to false alignments
and sequence homology artifacts. This is followed by variant annotation-based frequency, muta-
tion type, and other functional criteria. The combination fi ltering parameters along with inheri-
tance modeling (if available) is then performed to narrow the number of causative mutations to a
manageable number for further investigation. Knowledge databases (HGMD, OMIM) provide
functional information that is helpful for interpretation. In clinical laboratories, potential causative
variants are Sanger-confi rmed and a report is generated after review
Search WWH ::

Custom Search