Data from: Environmental metabarcodes for insects: in silico PCR reveals potential for taxonomic bias

Studies of insect assemblages are suited to the simultaneous DNA-based identification of multiple taxa known as metabarcoding. To obtain accurate estimates of diversity, metabarcoding markers ideally possess appropriate taxonomic coverage to avoid PCR-amplification bias, as well as sufficient sequence divergence to resolve species. We used in silico PCR to compare the taxonomic coverage and resolution of newly designed insect metabarcodes (targeting 16S) with that of existing markers (16S and COI) and then compared their efficiency in vitro. Existing metabarcoding primers amplified in silico less than 75% of insect species with complete mitochondrial genomes available, whereas new primers targeting 16S provided greater than 90% coverage. Furthermore, metabarcodes targeting COI appeared to introduce taxonomic PCR-amplification bias, typically amplifying a greater percentage of Lepidoptera and Diptera species, while failing to amplify certain orders in silico. To test whether bias predicted in silico was observed in vitro, we created an artificial DNA blend containing equal amounts of DNA from 14 species, representing 11 different insect orders and one arachnid. We PCR-amplified the blend using five primers sets, targeting either COI or 16S, with high-throughput amplicon sequencing yielding more than 6 million reads. In vitro results typically corresponded to in silico PCR predictions, with newly designed 16S primers detecting 11 insect taxa present, thus providing equivalent or better taxonomic coverage than COI metabarcodes. Our results demonstrate that in silico PCR is a useful tool for predicting taxonomic bias in mixed template PCR, and that researchers should be wary of potential bias when selecting metabarcoding markers.

昆虫群落研究可采用依托DNA同步鉴定多个类群的技术，即宏条形码（metabarcoding）。为获得准确的物种多样性评估结果，宏条形码标记应具备合适的类群覆盖度以规避PCR扩增偏好性，同时拥有足够的序列分化度以实现物种分辨。本研究采用电子PCR（in silico PCR）对比了新设计的靶向16S的昆虫宏条形码引物与现有标记（16S与COI）的类群覆盖度及分辨能力，并进一步通过体外（in vitro）实验验证其扩增效率。
现有宏条形码引物在电子PCR中仅能覆盖已公布完整线粒体基因组的昆虫物种中不足75%的类群，而新设计的靶向16S的引物覆盖度可达90%以上。此外，靶向COI的宏条形码似乎会引入类群特异性PCR扩增偏好性：通常可扩增出更高比例的鳞翅目（Lepidoptera）与双翅目（Diptera）物种，却无法在电子PCR中扩增出部分昆虫目级类群。
为验证电子PCR预测的扩增偏好性能否在体外实验中重现，本研究构建了人工混合DNA模板：包含14个物种的等量DNA，分别隶属于11个昆虫目与1个蛛形纲物种。我们采用5组靶向COI或16S的引物对该混合模板进行PCR扩增，并通过高通量扩增子测序（high-throughput amplicon sequencing）获得了超过600万条测序读段（reads）。
体外实验结果与电子PCR预测结果基本吻合：新设计的16S引物可检测出11个目标昆虫类群，其类群覆盖度与COI宏条形码相当甚至更优。
本研究结果表明，电子PCR可有效预测混合模板PCR中的类群扩增偏好性，同时提醒研究者在选择宏条形码标记时需警惕潜在的扩增偏差问题。