Miss Raisa Mikaelle Castillo1, Dr. Ambrocio Melvin Matias1
1Institute of Biology, University of the Philippines Diliman, Quezon City, Philippines
Biography:
Rai earned her Master of Science in Biology degree from the University of the Philippines Diliman. As a Project Technical Specialist, she works on detecting rare endemic fish species through eDNA analysis. Her primary research interests include population genetics and bioinformatics. When she isn't in the lab or coding on her pink mechanical keyboard, she is likely watching K-pop music videos or K-dramas.
Abstract:
Environmental DNA (eDNA) is a useful tool for profiling fish biodiversity and has proven to be more sensitive than traditional survey methods. Species detection makes use of classification pipelines that assign taxonomy to sequencing reads. Among commonly used pipelines are Kraken2, which classifies reads based on exact k-mer alignments, and Qiime2-Naïve Bayes Classifier, which establishes k-mer frequencies instead of exact alignment. However, these two have disagreements in their classification. Their performance also varies depending on the taxa being studied and reference database used. They also differ in accuracy to classify rare sequences that have no similarity to present databases. To evaluate the performance of the two classifiers, we extracted cytochrome c oxidase subunit I sequences of freshwater fishes with varying divergence time from Genbank. We deliberately excluded some species to simulate missing sequences in reference databases. Prior to classification, we added sequencing errors to the query reads. There were disagreements between classifications of the two pipelines, especially at the species level. Some rare sequences were correctly classified at higher taxonomic levels, but falsely classified at the species level. The two pipelines were then used to assess biodiversity of Lake Lanao, an ancient tropical lake that harbors Cyprinid fishes–most of which are considered extinct and have no reference sequences. There were some disagreements in the species-level assignment. Therefore, it is challenging to perform metabarcoding, due to its dependence on reference databases. Despite the analytical challenges involved, eDNA remains a valuable tool for providing insights into the biodiversity of Lake Lanao.