AQUADEX
eDNA / BIOINFORMATICS
Aquadex
Reading traces of life from deep-sea DNA. Classification meets novelty detection.
THE IDEA
Environmental DNA leaves clues about organisms that are difficult to observe directly. Aquadex connects sequence processing, taxonomic classification and novelty analysis in a reproducible workflow.
HOW IT WORKS
- 01Clean sequence reads
- 02Resolve sequence variants
- 03Classify taxonomy
- 04Compare embeddings
- 05Corroborate novelty
INSIDE THE BUILD
Nextflow coordinates processing stages and their dependencies.
Sequence quality control and variant resolution prepare the input for analysis.
Classification and embedding-based search offer complementary evidence.
Phylogenetic and protein-level checks help contextualise unfamiliar sequences.
An unfamiliar sequence is a hypothesis to investigate, not automatic proof of a new species.
THE TRADE-OFFS
Results depend on reference database coverage and which models are available. Deep-sea taxa are underrepresented; missing references can resemble biological novelty.
BUILT WITH
- Python
- Nextflow
- FastAPI
- FAISS
- Docker