Validation and optimization of a rare-taxon filtering algorithm
Rare taxa filtering amplicon sequencing-based microbiome studies constitutes one of the most critical yet least standardized methodological decisions, directly impacting diversity estimation and ecological interpretation. Conventional approaches tend to apply arbitrary thresholds based on abundance...
| Autor: | |
|---|---|
| Tipo de recurso: | tesis de maestría |
| Fecha de publicación: | 2026 |
| País: | España |
| Institución: | Universitat Oberta de Catalunya (UOC) |
| Repositorio: | O2, repositorio institucional de la UOC |
| OAI Identifier: | oai:dnet:orepositorio::5f567dace26fa4ac3003c69894a861ca |
| Acceso en línea: | https://hdl.handle.net/10609/155216 |
| Access Level: | acceso abierto |
| Palabra clave: | rare taxa SynCom ASV OTU Bioinformatics -- FMDP Bioinformàtica -- TFM |
| Sumario: | Rare taxa filtering amplicon sequencing-based microbiome studies constitutes one of the most critical yet least standardized methodological decisions, directly impacting diversity estimation and ecological interpretation. Conventional approaches tend to apply arbitrary thresholds based on abundance or other parameters, leading to either the removal of biologically relevant taxa or the inflation of diversity due to technical noise. This study presents the validation and optimization of a quantitative algorithm for rare taxa filtering, grounded in the weighting of alpha and beta diversity metrics with distinct weight distributions. The methodology focuses on determining optimal k thresholds using synthetic communities (SynCom) of known composition as an objective reference (ground truth). The workflow incorporates an iterative ASV reintroduction procedure (add-one pipeline) and evaluates performance using classification metrics, while also comparing the behavior of different diversity metrics (such as Chao1, Jaccard, or Aitchison). Finally, the approach is validated on real environmental data. The results demonstrate that the algorithm maximizes the harmonic mean of precision and recall (F1 – score), converging at a controlled False Positive Rate (FPR) of ~0.2. Unlike conventional static filters, this strategy preserves the robustness of the core community while correcting overestimation on the rare biosphere, offering a reproducible solution that is more faithful to biological reality. |
|---|