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...

Descripción completa

Detalles Bibliográficos
Autor: Balboa Ortega, Antonio Jesús
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
Descripción
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.