Exome sequencing for gene discovery: time does not stand still.
Exome sequencing for gene discovery: time does not stand still.
复制标题
用于基因发现的外显子组测序：时间不会静止。
DOI：
10.1002/ana.23660
复制
发表时间：
2012
影响因子：
11.2
通讯作者：
LeDoux,MarkS
中科院分区：
文献类型：
--
作者：
LeDoux,MarkS
全外显子组测序的快速发展和不断完善使得建立标准的尝试几乎是徒劳的。此外，在不久的将来，全基因组测序可能会完全取代全外显子组测序。对于所有高通量DNA测序技术，需要强调的是，读取精度比读取深度重要得多。即使是深度读取，读取错误也并不少见。从理论上讲，如果所有的读取都是长、无偏见和100%准确的，那么只需要很浅的覆盖。此时，覆盖范围取决于多种因素，包括外显子组捕获方法，测序平台，平均读取长度，覆盖范围定义，外显子组定义（CCDS+/−RefSeq,+/−GENCODE）和每次运行分析的不同DNA样本数量。在许多情况下，超深度覆盖的成本高得令人望而却步，并且不能消除假阳性和假阴性。鉴于下一代测序技术的已知局限性，我们使用gDNA和cDNA的Sanger测序（正向和反向）来排除我们的宫颈肌张力障碍亲属的编码和非编码（非翻译区，内含子和启动子）TOR1A突变。用gDNA qPCR排除拷贝数变异。在dbSNP、Exome Variant Server和1000个基因组中报告的很大一部分变异没有通过Sanger测序验证，可能是读取错误。例如，Exome Variant Server报告了TOR1A（一个与早发性全身性肌张力障碍相关的5外显子基因）的11个错义变异，这些变异被预测可能或可能被PolyPhen破坏，所有这些变异的平均样本读取深度为80X。另一方面，考虑到常染色体显性肌张力障碍（SGCE、GCH1、TOR1A、CIZ1和THAP1）的外显率和人群患病率降低，在假定正常对照的大型数据库中发现潜在的因果变异也就不足为奇了。
The rapid evolution and continued refinement of whole-exome sequencing renders attempts to establish standards almost futile. Moreover, whole-genome sequencing may entirely replace whole-exome sequencing in the very near future. For all high-throughput DNA sequencing technologies, it should be emphasized that read accuracy is much more important than read depth. Even with deep reads, read errors are not uncommon. Theoretically, only shallow coverage would be needed if all reads were long, unbiased and 100% accurate. At this time, coverage depends on a variety factors including method of exome capture, sequencing platform, average read length, definition of coverage, definition of the exome (CCDS+/− RefSeq,+/− GENCODE) and the number of distinct DNA samples analyzed per run. Ultra-deep coverage is prohibitively expensive in many situations and does not eliminate false positives and false negatives. Given the known limitations of nextgeneration sequencing technologies, Sanger sequencing (forward and reverse directions) of gDNA and cDNA was used to confidently exclude coding and non-coding (untranslated regions, introns and promoter) TOR1A mutations in our kindred with cervical dystonia. 1 Copy number variants were excluded with qPCR of gDNA.A significant proportion of the variants reported in dbSNP, Exome Variant Server, and 1000 Genomes have not been validated with Sanger sequencing and are probably read errors. For example, Exome Variant Server reports 11missense variants in TOR1A (a 5 exon gene associated with early-onset generalized dystonia) that are predicted to be possibly or probably damaging by PolyPhen and average sample read depth was> 80X for all of these variants. On the other hand, it would not be surprising to find potentially causal variants in large databases of putatively normal controls given the reduced penetrance and population prevalence of autosomal-dominant dystonia (SGCE, GCH1, TOR1A, CIZ1 and THAP1).