Overview

This project is my senior honors thesis at the University of Texas at Austin, conducted in the DiFiore Lab (Dec 2024 – Present). It encompasses two tightly coupled components: a computational pipeline that processes raw Nanopore sequencing data from 83 environmental soil samples (~6 million reads) into reliable mammalian species identifications, and an interactive visualization platform that makes those results accessible to both scientific and public audiences.

Methods / Approach

Pipeline — End-to-end Nanopore processing:

• Collected environmental DNA (eDNA) soil samples from field sites across the Magdalena River Valley drainage basin, Colombia.
• Ran preprocessing and QC: adapter trimming (Cutadapt), quality filtering (NanoFilt), and run-level QC reporting (MultiQC).
• Performed consensus clustering via NGSpeciesID to correct random ONT base-call errors, achieving >99% consensus accuracy across 12S and 16S rRNA marker regions.
• Assigned taxonomy via megaBLAST against a custom mammalian reference library, with species identification and conservation status lookup through the GBIF Metabarcoding Data Toolkit API, NCBI, and IUCN.

Visualization platform:

• Developed a multi-audience web application layering sequencing results onto ArcGIS-hosted species distribution data, with separate views optimized for researchers and the general public.
• Platform allows users to explore species detections, geographic distributions, and IUCN conservation status interactively across the Magdalena River Valley corridor.

Key Results

• Processed 83 soil samples and ~6 million raw Nanopore reads through a fully automated end-to-end pipeline.
• Achieved >99% consensus accuracy using NGSpeciesID-based clustering on 12S/16S rRNA amplicons.
• Successfully identified mammalian taxa including several species of conservation concern in the Magdalena Valley.
• Platform demonstrates the full arc: raw ONT reads → consensus sequences → taxonomic assignments → interactive visualization, bridging bioinformatic outputs and actionable conservation insight.

Tools & Technologies

• ONT MinION / Nanopore — long-read sequencing platform
• Cutadapt, NanoFilt, MultiQC — adapter trimming, quality filtering, QC reporting
• NGSpeciesID — consensus clustering for ONT error correction
• megaBLAST / NCBI — taxonomic identification from consensus sequences
• GBIF Metabarcoding Data Toolkit API — species occurrence and range data integration
• IUCN — conservation status lookup
• ArcGIS — geospatial species-range data and map layers
• Bash / Shell — pipeline automation and HPC job submission (SLURM)