Simple approaches for evaluation of OTU quality based on dissimilarity arrays

Autor: Marie-Josée Cros, Jean-Marc Frigerio, Nathalie Peyrard, Alain Franc
Jazyk: angličtina
Rok vydání: 2024
Předmět:
Zdroj: Metabarcoding and Metagenomics, Vol 8, Iss , Pp 25-44 (2024)
Druh dokumentu: article
ISSN: 2534-9708
DOI: 10.3897/mbmg.8.108649
Popis: An accurate and complete taxonomic description of the diversity present in an environmental sample is out of reach at this time. Instead, metabarcoding is used today and it is expected that OTUs represent a category relevant for biodiversity inventories on a molecular basis. However, artefacts in the production of OTUs can occur at different stages and may impact ecological conclusions. We propose to evaluate the quality of OTUs in a sample by characterising the deviation of each OTU’s dissimilarity array from that of an ideal OTU where all sequences are at distances smaller than the barcoding gap. We consider two deviations: the creation of composed OTUs, corresponding to the artificial merging of several OTUs and the creation of noisy OTUs that contain some sequences that are loosely associated with the core sequence of the OTUs and that do not form a compact subgroup. We propose a simple and automatic 2-step method that successively categorises the OTUs of a sample as composed or single and then identifies OTUs with noise amongst the single ones. The associated code is available at https://forgemia.inra.fr/alain.franc/otu_shape. We applied the method on 32 samples of diatoms from Arcachon Bay (France) that represent contrasted environmental conditions and we obtained good agreement with expert categorisation of OTUs. We suggest that single OTUs without noise can be used as such for further ecological studies. Composed OTUs should be post-treated with classical clustering or community detection tools. The quality of single OTUs with noise remains to be further tested via supplementary studies on a diversity of organisms.
Databáze: Directory of Open Access Journals