eDNAssay: A machine learning tool that accurately predicts qPCR cross‐amplification

Abstract

Environmental DNA (eDNA) sampling is a highly sensitive and cost‐effective technique for wildlife monitoring, notably through the use of qPCR assays. However, it can be difficult to ensure assay specificity when many closely related species co‐occur. In theory, specificity may be assessed in silico by determining whether assay oligonucleotides have enough base‐pair mismatches with nontarget sequences to preclude amplification. However, the mismatch qualities required are poorly understood, making in silico assessments difficult and often necessitating extensive in vitro testing—typically the greatest bottleneck in assay development. Increasing the accuracy of in silico assessments would therefore streamline the assay development process. In this study, we paired 10 qPCR assays with 82 synthetic gene fragments for 530 specificity tests using SYBR Green intercalating dye (n = 262) and TaqMan hydrolysis probes (n = 268). Test results were used to train random forest classifiers to predict amplification. The primer‐only model (SYBR Green results) and full‐assay model (TaqMan probe‐based results) were 99.6% and 100% accurate, respectively, in cross‐validation. We further assessed model performance using six independent assays not used in model training. In these tests the primer‐only model was 92.4% accurate (n = 119) and the full‐assay model was 96.5% accurate (n = 144). The high performance achieved by these models makes it possible for eDNA practitioners to more quickly and confidently develop assays specific to the intended target. Practitioners can access the full‐assay model online via eDNAssay (https://nationalgenomicscenter.shinyapps.io/eDNAssay), a user‐friendly tool for predicting qPCR cross‐amplification.

Validation of a nutria (Myocastor coypus) environmental DNA assay highlights considerations for sampling methodology

“…We identified ≥16 total mismatches between assay oligonucleotides and sequences from co-occurring nontarget vertebrate species. This ought to be more than enough mismatches to preclude amplification (Kronenberger et al, 2022), suggesting that the assay is highly nutria specific. Furthermore, we did not identify any additional potential nontarget taxa that occur within the United States with our Primer-BLAST search of the GenBank database.…”

Molecular identification and environmental DNA detection of gill lice ectoparasites associated with Brook Trout declines

“…We further tested for potential qPCR cross‐amplification with the eDNAssay AI Web tool (Kronenberger et al. 2022) by using the Lernaeopodidae sequence alignment as data input. Any species with assignment probabilities greater than 0.29 were then subjected to follow‐up in vitro specificity testing, as recommended by Kronenberger et al.…”

“…Assay specificity was validated in silico against the full NCBI Nucleotide (nr/nt) database with NCBI Primer-BLAST (Basic Local Alignment Search Tool; Ye et al 2012) using forward and reverse primers with the following parameters: nontarget matches with more than nine mismatches or amplicons greater than 3000 bp were ignored; no organisms were excluded; the maximum number of sequences returned by BLAST was 100,000; the BLAST expect I value was 100,000; the BLAST word size was 6; and all other parameters were left as default. We further tested for potential qPCR cross-amplification with the eD-NAssay AI Web tool (Kronenberger et al 2022) by using the Lernaeopodidae sequence alignment as data input. Any species with assignment probabilities greater than 0.29 were then subjected to follow-up in vitro specificity testing, as recommended by Kronenberger et al (2022), by performing qPCR using gBlocks synthetic DNA fragments (Integrated DNA Technologies) that were designed to match nontarget COI haplotypes (Table 1).…”

“…We further tested for potential qPCR cross-amplification with the eD-NAssay AI Web tool (Kronenberger et al 2022) by using the Lernaeopodidae sequence alignment as data input. Any species with assignment probabilities greater than 0.29 were then subjected to follow-up in vitro specificity testing, as recommended by Kronenberger et al (2022), by performing qPCR using gBlocks synthetic DNA fragments (Integrated DNA Technologies) that were designed to match nontarget COI haplotypes (Table 1).…”

Enabling robust environmental DNA assay design with “unikseq” for the identification of taxon‐specific regions within whole mitochondrial genomes

“…Many aspects of assay design can benefit from emerging technological approaches. For example, machine learning has been employed to accurately predict quantitative real‐time polymerase chain reaction (qPCR) cross‐amplification in barcode regions (Kronenberger et al, 2022), and k‐ mer‐based bioinformatics approaches have been used to predict phenotypes such as antibiotic resistance from sequencing data (Aun et al, 2018; Nurmukanova et al, 2022; Simpson et al, 2009; Warren et al, 2019, 2015). One of the most impactful developments has been an increase in sequencing data availability from a large range of taxa that constitute the targets of monitoring efforts as well as confounding taxa.…”