No abstract

Changes in the macroinvertebrate community were investigated over 10 months at four sites along a 19 km salinity gradient (0.21-2.60 g l )1 ) in a sixth-order stream, the Meurthe River, northeastern France. Abiotic characteristics other than salinity were similar between the sites. Macroinvertebrate taxonomic richness decreased by 30% downstream of the 1.4 g l )1 sites while diversity, evenness or total abundance of taxa did not change along the gradient. In terms of functioning, a slight change in relative abundances of invertebrate feeding groups followed the salinity gradient. Eight invertebrate assemblages occurred within specific salinity distributions were identified. The exotics Gammarus tigrinus, Dreissena polymorpha, Corbicula fluminalis and Corophium curvispinum, were more abundant at the highest salinity site. These results suggest that rising salinity concentrations drastically affect the species composition, including favouring exotic species.

Cytochrome c oxidase I (COI) is a powerful marker for DNA barcoding of animals, with good taxonomic resolution and a large reference database. However, when used for DNA metabarcoding, estimation of taxa abundances and species detection are limited due to primer bias caused by highly variable primer binding sites across the COI gene. Therefore, we explored the ability of the 16S ribosomal DNA gene as an alternative metabarcoding marker for species level assessments. Ten bulk samples, each containing equal amounts of tissue from 52 freshwater invertebrate taxa, were sequenced with the Illumina NextSeq 500 system. The 16S primers amplified three more insect species than the Folmer COI primers and amplified more equally, probably due to decreased primer bias. Estimation of biomass might be less biased with 16S than with COI, although variation in read abundances of two orders of magnitudes is still observed. According to these results, the marker choice depends on the scientific question. If the goal is to obtain a taxonomic identification at the species level, then COI is more appropriate due to established reference databases and known taxonomic resolution of this marker, knowing that a greater proportion of insects will be missed using COI Folmer primers. If the goal is to obtain a more comprehensive survey the 16S marker, which requires building a local reference database, or optimised degenerated COI primers could be more appropriate.

Cytochrome c oxidase I (COI) is a powerful marker for DNA barcoding of animals, with good taxonomic resolution and a large reference database. However, when used for DNA metabarcoding, estimation of taxa abundances and species detection are limited due to primer bias caused by highly variable primer binding sites across the COI gene. Therefore, we explored the ability of the 16S ribosomal DNA gene as an alternative metabarcoding marker for species level assessments. Ten bulk samples, each containing equal amounts of tissue from 52 freshwater invertebrate taxa, were sequenced with the Illumina NextSeq 500 system. In comparison to COI, the 16S marker amplified more insect species and amplified more equally, probably due to decreased primer bias. Rough estimation of biomass might thus be less biased with 16S than with COI. According to these results, the marker choice depends on the scientific question. If the goal is to obtain a taxonomic identification at the species level, then COI is more appropriate due to established reference databases and known taxonomic resolution of this marker, knowing that a greater proportion of species will be missed using COI Folmer primers. If the goal is to obtain a more comprehensive survey in a context where it is possible to build a local reference database, the 16S marker could be more appropriate.