Revelando a variação da comunidade de peixes associada ao riverscape através do método de metabarcoding de DNA ambiental

• Ano de defesa: 2025
• Autor(a) principal: López Paría, Ricardo Jesús
• Orientador(a): Galetti Junior, Pedro Manoel
• Tipo de documento: Dissertação
• Tipo de acesso: Acesso aberto
• Idioma: por
• Instituição de defesa: Universidade Federal de São Carlos; Câmpus São Carlos
• Programa de Pós-Graduação: Programa de Pós-Graduação em Ecologia e Recursos Naturais - PPGERN
• Palavras-chave em Português: Análises multivariadas; Riverscape; Uso do solo; eDNA; Ecologia; Ictiofauna; Neotrópico
• Palavras-chave em Inglês: Metabarcoding; Multivariate analyses; Land use; Ecology; Ichthyofauna; Neotropics; Environmental DNA
• Área do conhecimento CNPq: CIENCIAS BIOLOGICAS::ECOLOGIA; CIENCIAS BIOLOGICAS::ECOLOGIA::ECOLOGIA APLICADA
• Link de acesso: https://hdl.handle.net/20.500.14289/22130

Anthropic land use significantly impacts aquatic biodiversity by influencing the structure and composition of fish communities and altering freshwater ecosystems. In this study, we investigated how land use and other environmental variables influence the structure and composition of fish communities in the Mundaú River Basin, located in the Pernambuco Endemism Center (PEC), using the eDNA metabarcoding approach. Water samples were collected from headwater streams distributed across four land use categories (Forest, Agriculture, Pasture, and Urban) and analyzed through amplification of the mitochondrial 12S region using MiFish-U primers. Sequencing generated 3,468,152 raw reads, which, after bioinformatic processing and quality filtering, resulted in a final dataset with 32,940 reads and 39 MOTUs assigned to 36 nominal species or taxa belonging to six fish orders. Among these, two taxa, Oreochromis niloticus and Coptodon rendalli, were classified as exotic species. The species accumulation curve showed a reduction in the discovery rate in the final samples, suggesting that the sampling effort was sufficient to capture most of the local diversity. Species richness ranged from 2 to 17 species per sampling point. The Shannon-Wiener diversity index (H') varied from 0.076 to 2.099, with higher values in Forest, Agriculture, and Pasture areas, while urban points exhibited greater variability and lower mean diversity values. The Kruskal-Wallis test indicated significant differences in Shannon-Wiener diversity (χ² = 9.3224; p = 0.0253) among land use types. The post-hoc Mann-Whitney test with Benjamini-Hochberg correction revealed that the only statistically significant difference was between Agriculture and Urban areas (adjusted p = 0.049), while the comparison between Forest and Urban lost significance after adjustment (p = 0.031; adjusted p = 0.061). PERMANOVA analysis indicated that land use explains 17.2% of the variation in fish community composition (F = 1.38; p = 0.096), representing moderate evidence of anthropogenic influence, although without complete segregation between environments, as shown by the NMDS overlap. Pairwise comparisons revealed that the most pronounced differences occurred between Forest and Urban (p = 0.071; adjusted p = 0.297) and between Agriculture and Urban (p = 0.099; adjusted p = 0.297). The Indicator Species Analysis (IndVal) identified Callichthys callichthys and Poecilia sp.1 as indicators of pasture areas, while Oreochromis niloticus showed a strong association with urban environments. No indicator species were detected for Forest or Agriculture areas, likely due to the predominance of generalist species and the presence of the invasive species Coptodon rendalli in both environments. The SIMPER analysis revealed that dissimilarity between environments was mainly driven by Coptodon rendalli, Oreochromis niloticus, Poecilia sp.1, and Hoplias sp.2. Principal Component Analysis (PCA) identified two main axes that explained 45.22% of the total variance in environmental and landscape data. PC1 (25.64%) reflected variation associated with local hydrological characteristics, while PC2 (19.58%) represented gradients of land use. The most relevant environmental variables included Mean Depth (PC1: -0.391), the percentage of pasture land in the drainage area (PC1: -0.386), and forest cover in the riparian buffer (PC2: 0.502). Redundancy Analysis (RDA) showed that the proportion of urban area in the drainage (F = 3.82; p = 0.002) was the most predictive variable of fish community structure, followed by the presence of leaf litter in the substrate (F = 2.51; p = 0.023). Together, these variables explained 23.16% of the total variation in MOTU composition. This suggests that urbanization has the greatest impact on fish community structure in the basin, while the presence of structured substrates like leaf litter plays a secondary role in maintaining biodiversity. The results of this study demonstrate that riverscape heterogeneity combined with land use plays a crucial role in determining diversity patterns and structuring fish communities. Furthermore, the findings reinforce the potential of eDNA metabarcoding as a sensitive and effective tool for ecological monitoring, especially in megadiverse regions with a lack of taxonomic inventories such as the MNCE. These findings provide valuable insights for conservation and management strategies, highlighting the need to mitigate the effects of urbanization and invasive species introductions, as well as the importance of strengthening regional genetic reference databases to improve future eDNA-based applications in aquatic ecosystem management.