Cross-contamination risk
During the metabarcoding PCR (here specified as the PCR in which the metabarcoding marker is targeted), relatively short DNA sequences (typically <300 bp) are enriched through amplification. Especially when targeting trace amounts of DNA, PCR amplification can be highly susceptible to contamination and thereby to false positives. The risk of contamination when preparing metabarcoding PCRs is the same no matter which of the three overall metabarcoding approaches is used. Moreover, regardless of the metabarcoding strategy employed, cross-contamination can happen between nucleotide tagged and indexed primer stocks (which are delivered at very high molarity). The risk of this happening will be similar between the strategies and will depend on the number of samples and the chosen setup within the employed strategy. In the following, rather than discussing primer contamination, we will focus on how the three main metabarcoding approaches differ in risk of cross-contamination between PCR products after the metabarcoding PCR.
In the one-step PCR approach (Fig. 2B) and the tagged PCR approach (Fig. 2D), PCR products are labelled during the metabarcoding PCR amplification. In the one-step PCR approach, the metabarcoding PCR is carried out with primers that target the selected marker and carry both sequencing adapters and indices. This way, the indexed PCR products can be immediately sequenced following this one PCR step (Fig. 2B). If the indexed ready-to-sequence libraries are to be pooled into one pool before sequencing, then cross-contamination between indexed amplicon libraries is obviously not of concern. However, if more sequencing pools are made in which the same index combinations occur across multiple samples, then cross-contamination between the sequencing pools can be an issue. A solution is to process them in separate sequencing run batches to avoid cross-contamination. In the tagged PCR approach, amplicons will be 5’ nucleotide tagged following the metabarcoding PCR, which means that cross-contamination between tagged PCR products is not of concern. However, until the amplicon pools are indexed during library preparation there is a risk of cross-contamination between amplicon pools if the same tag combinations are used in different amplicon pools (Schnell et al.2015). Some laboratories do not reuse tag-primer combinations to further reduce contamination risk (see Murrayet al. 2015).
In the two-step approach, sample-specific labelling is not carried out during the metabarcoding PCR. This creates a risk of cross-contamination between unlabelled PCR products when handling them prior to the second PCR (Zizka et al.2019). Therefore, this metabarcoding approach has the greatest theoretical risk of cross-contamination between PCR amplicons (Fig. 2C). It is worth mentioning that a few studies adopt modifications of the two-step approach that eliminates this kind of cross-contamination. One is to include nucleotide labelling in the first PCR, see Kitson et al. (2018), and the other is to carry out both of the two PCRs, i.e. to include both two primer sets, in the same reaction, see for example Clarke et al. (2014a).
Irrespective of the chosen approach, cross-contamination can be detected and filtered out by including sample replicates, PCR replicates, and positive and negative controls. Thus, these should be included in the laboratory workflow and sequencing (e.g. Bistaet al. 2017). An important measure that enables one to filter out potential contamination during data processing is to use different nucleotide tag and/or library index combinations on each sample’s individual PCR replicates as this will allow for restrictive sequence processing across each sample’s PCR replicates (Alberdi et al.2018).