Methods
Summary
This project consist in tree main steps to cover environmental DNA field collection to the construction of DNA libraries in our laboratory:
Environmental DNA collection and DNA extraction. The use of eDNA to detect species is considerably more challenging in marine environments owing to greater dilution, increased mixing, and higher salinity, thus, we propose adjustments accordingly.
To capture changes in taxonomic composition, the collection of environmental DNA (eDNA) will be performed in a representative month of each season of the year covering at least two consecutive years. Environmental DNA samples will be taken in the “Santa Rosa'' estuary located in Sonora, Mexico. This site is near the northernmost limit of mangroves in the Eastern Pacific and exhibits remarkable changes in temperature and salinity. Also, there are three of the four most representative species of mangroves in Mexico: Avicennia germinans, Rhizophora mangle, and Laguncularia racemosa.
For each mangrove species, soil, estuarine water, leaves and tree surface (bark and primary branches) samples will be taken. Estuarine water sample will filter across a Millipore 0.2 μm hydrophilic nylon membrane (Merck Millipore) using a peristaltic pump. The membrane disc containing eDNA and cellular material from the water will be conserved in liquid nitrogen; for each procedure, sterile deionized water will also be filtered in the same way as environmental DNA samples to discard contamination in the sample collection. Because turbid estuarine waters characteristic of mangroves have high concentrations of suspended organic and inorganic particles, it is probable that filters become clogged. Thus, together with filtered water, we immediately store non-filtered estuarine water in liquid nitrogen until DNA extraction. By “washing leaves”, we will resuspend in esteril solution any trace of eDNA to subsequently store in liquid nitrogen. Tree surface samples will be taken following Valentin et al. (2020). Three replicates of each type of eDNA sample will be taken, separated by a minimum of 200 m trying to cover the greatest representativeness of the estuary. Before the DNA extraction step, we will use for all eDNA samples the nucleoSpin Inhibitor Removal Kit for removal of PCR inhibitors. Subsequently, DNA extraction will be performed following the Qiagen DNeasy Blood & Tissue Kit protocol considering each environmental sample and two quality controls: field contamination tests and negative control during the extraction procedure. Another option to overcome PCR inhibitors is to implement the standard protocol phenol-chloroform-K proteinase and then purify DNA using DNA Purification Magnetic Beads to increase the PCR amplification success. In both cases, DNA quantification will be performed with Qubit dsDNA HS Assay.
DNA Amplification. Due to the wide diversity of taxa that hold this ecosystem, first we will get started with the most popular vertebrate taxa employing eDNA, the fishes. Please note that screening the biodiversity living in mangrove forest represents a methodological challenge, thus, at the time we assess fishes diversity, we must ensure the successful amplification of other less studied groups, such as birds, mammals, reptiles and invertebrate taxa using soil, estuarine water, tree surface and leaf samples.
To amplify eDNA-fishes fragments, we will use a two-step PCR amplification method including the extracted eDNA samples, a negative control and two positive ones: DNA samples from a known fish species and another consisting of a mixture of DNA from several known species. In the first amplification, we will use a specific primer flanking a hypervariable mitochondrial region of the 12S rRNA gene (Miya et al., 2015) with an average length of 170bp using the reported amplification conditions. Amplifications will be performed with a high-fidelity Taq polymerase enzyme. After, the PCR products will be visualized in a 1.5% agarose gel to verify the amplification of the environmental DNA samples and positive controls and the non-amplification of the negative ones. In the second amplification, the same primers will be used but labeled at the 5' end (in the forward and reverse directions) to allow the copying of multiple sequences at the same time. For amplification of the fragment, the product of the first PCR will be used as a template. They will later be visualized and quantified.
DNA libraries The libraries will be prepared using approximately 250 ng of product from the second PCR, mixing all the samples in the same pool. The libraries will have unique identifiers of approximately 6pb, as well as the barcodes of the Illumina adapters. The quality of the resulting libraries will be evaluated with ultra-precision equipment (a tapestation or Bioanalizer); their preparation and sequencing will be performed with a MiSeq equipment. The bioinformatics analysis of the first experiments will be carried out with the Unix system; filtering sequences based on their quality (fastq) and demultiplexing based on the presence of both directions of the sequence. Molecular Operational Taxonomic Units (MOTU) will be formed by grouping the sequences using a radius of 1; MOTUs with frequencies <0.005% will be eliminated. The MOTUs will be annotated by comparison with the nucleotide (nt) database of the National Center for Biotechnology Information (NCBI) using BLAST; using a percentage of identity higher than that proposed by the authors who developed the primers (97%) and those used in other studies with metabarcoding. Finally, to annotate a taxonomic category for those sequences with equal identity percentages, we will use the following taxonomic category in common, that is, if two sequences with 98% identity coincide with two species of the same genus, the annotated unit will be at the genus level. At the end, the annotations will be organized by each mangrove species and seasonality. In this way, it will be possible to infer seasonal patterns due to the absence and presence of MOTUs in different seasons and due to the effect of the zonation of mangroves, which is mediated by the influence of environmental variables and tides.
Challenges
Our first challenge is the obtention of eDNA (extraction) because in marine environments could be more difficult to capture fragments of DNA due to high salinity and secondary metabolites produced by mangroves. So, we proposed here modified methods of sample collection and DNA extraction to improve success in these first step of the project.
The second challenge is the prediction of species interaction networks based on machine learning because this new approach makes a lot of assumptions that is likely that our data will not be appropriated to be informative to get confidential results. However, we have already considered other simpler methods to construct/infer interaction networks.
Pre Analysis Plan
Data analyses To evaluate the differences in taxonomic composition between mangrove species and seasonality, analysis of variance (ANOVA) will be used. The assumptions of normality (Kolmogorov Smirnov) and heteroscedasticity will be tested prior to the analysis of variance. When the ANOVA suggests significant differences, the Tukey-Kramer significant differences (HSD) post hoc test will be used. Taxonomic composition diversity will be estimated using “true diversity index” corrections using the non-parametric Chao1 estimator to represent 0th order diversity (taxa richness in this case). To infer changes in the trophic chain throughout the year, the assignment of trophic levels and any other characteristics of the life history of the taxonomic annotations will be carried out through the review of the literature using specialized databases as sources of information such as NCBI, the world register of Marine Organism, Encyclopedia of Life, Integrated Taxonomic Information System, and peer-reviewed articles. This procedure will be carried out by two different people to confirm the assigned trophic group (e.g. primary or secondary consumers). To evaluate the influence of environmental variables on the seasonal variation of taxonomic richness, environmental data layers from Bio-Oracle will be obtained such as temperature, salinity, primary productivity, chlorophyll, among others. The correlation between explanatory variables will be evaluated to avoid important biases in the subsequent analysis. Those with an index greater than 0.7 will be considered highly correlated variables. To identify relationships between taxonomic richness and habitat heterogeneity, generalized linear models will be used. The taxonomic richness index will be the response variable and the habitat heterogeneity, represented by the environmental variables will be the explanatory variables. The Akaike Information Criterion will be used to select the best model. Finally, we will follow Strydom et al. (2022) to predict species interaction networks based on machine learning; and model changes in biodiversity composition in the face of future climate circulation models
Protocols
This project has not yet shared any protocols.
