Automated DNA-based plant identification for large-scale biodiversity assessment

Area of Science:

• Ecology
• Genetics
• Bioinformatics

Background:

• Tropical forest degradation necessitates improved large-scale biodiversity assessment methods.
• Current DNA barcoding identification methods face challenges with incomplete reference databases in hyperdiverse ecosystems.
• Phylogenetic methods are robust but time-consuming, and multiple alignment of variable markers is difficult.

Purpose of the Study:

• To develop and validate a pipeline combining phylogenetic and phenetic methods for DNA-based taxonomic assignment.
• To address the limitations of existing methods for hyperdiverse tropical ecosystems with sparse reference data.
• To create a tool for efficient, high-throughput biodiversity assessment in large-scale projects.

Main Methods:

• Developed a pipeline (BAGpipe) integrating database construction, sequence clustering, distance calculation, multiple alignment, and phylogenetic inference.
• Optimized parameters for the psbA-trnH marker.
• Tested the pipeline on plant samples and herbivore diet sequences from Nicaraguan seasonally dry forests using a local reference database.

Main Results:

• The pipeline proved efficient and reliable for high-throughput taxonomic assignment.
• Strict tree-based criteria were crucial for avoiding false positives in species identification.
• The approach demonstrated effectiveness even with incomplete local reference databases.

Conclusions:

• The combined phylogenetic and phenetic approach is a robust solution for DNA-based biodiversity assessment.
• The BAGpipe pipeline facilitates accurate and efficient species identification in large-scale biodiversity projects.
• The methodology is adaptable to various markers and taxa, supporting conservation efforts.