Stream: genomics
Topic: Sequence.readDepth
Kevin Power (Oct 08 2018 at 18:43):
At the WGM, we discussed the following tracker:
13834 / Add seqeunce.readdepth
We current have Sequence.readCoverage, which is defined as "Coverage (read depth or depth) is the average number of reads representing a given nucleotide in the reconstructed sequence." It seems that readCoverage is an average depth for the whole test, where perhaps readDepth is usually more at a specific location? But, we lacked a real expert in the room at the WGM, so anyone else is welcome to weigh in.
Dora Walter (Oct 09 2018 at 09:07):
Quote from our NGS Example Laboratory Findings
Genetic data evaluation:
The sensitivity of the test is depending on the tumor load, the sample quality and sequencing depth. A coverage of 120 reads per base achieves a theoretical sensitivity of 99% for the detection of variants with a ≥ 10% NAF. In this case, 95.1% of the targeted regions were covered by a minimum of 120 high-quality sequencing reads per base. Variants are named according to the HGVS recommendations without any information regarding the cis or trans configuration.
--> It seems to me that both terms are used identically here
Joel Schneider (Oct 10 2018 at 10:16):
The Sequence.readCoverage definition says average depth, and the sequencing depth definition says minimum depth.
If average depth is useful, it may be best to specify the type of average (e.g. mean, mode, median) and possibly also a standard deviation.
Gideon Giacomelli (Oct 10 2018 at 15:03):
what about using 82121-5 Allelic read depth as a name for a coverage for SNPs and small indels (<=30bps)
https://r.details.loinc.org/LOINC/82121-5.html?sections=Comprehensive
Gideon Giacomelli (Oct 10 2018 at 15:09):
coverage indeed is not precisely defined and there is also an ongoing discussion if the average or the median should be used. IMO using the median gives a better idea about the quality of the data especially if whole genome sequencing was used
Andrea Pitkus, PhD, MLS(ASCP)CM, CSM (Oct 11 2018 at 14:38):
What are your laboratories actually reporting? That would best help inform the shape of this item.
Kevin Power (Oct 11 2018 at 15:54):
Here are a few statements from example reports:
Exomes are sequenced to an average target coverage of 125X with 90-95% of bases sequenced to at least 20X coverage. All genes relevant to the indication had adequate coverage (>95% at 20X). The table below provides the list of these genes (and their coverage at ≥20X). <list of genes, each with a "20x coverage (%)" value>
>97% of target bases are covered at >20x, mean coverage of target bases >120x.
The Color test has 100% coverage for all regions in our reportable range >20X. Our median cov- erage across our samples is >250X (can exceed 1000X) and our minimum acceptance criteria for depth is: >99% at >50X and 100% at 20X. Any exceptions to this are noted in the Limitations section.
Last updated: Apr 12 2022 at 19:14 UTC
