DNA reference database for macroinvertebrates

This repository hosts a ready-to-use database for the study of macroinvertebrate diversity using DNA metabarcoding with the fwhF2/EPTDr2n primer set (Leese et al. 2022).

The database was developed from the MIDORI database and completed with some data from the BOLD database.

The process of preparation and curation of the database is described below:

1. The MIDORI CO1 RAW v247 database was downloaded from the official servers.

2. Reference sequences were matched against the fwhF2 forward primer sequences. Non-matching sequences were removed. Bases preceding the forward primer were removed.

3. Reference sequences with a length of less than 142 bp were removed. Bases following the position 142 were removed.

4. Taxonomy was simplified to keep only 8 ranks: superkingdom, kingdom, phylum, class, order, family, genus and species.

5. Taxonomic nomenclature was harmonized using `refdb::refdb_clean_tax_harmonize_nomenclature`

6. Extra words were removed from taxonomic names using `refdb::refdb_clean_tax_remove_extra`

7. Subspecific information were removed from taxonomic names using `refdb::refdb_clean_tax_remove_subsp`

8. Missing taxonomic names were identified and normalized using `refdb::refdb_clean_tax_NA`.

9. Hybrids and taxonomic names with qualifiers of uncertainty were converted to NA.

10. Duplicates (identical sequences and taxonomy) were removed.

11. Sequences with more than ambiguous nucleotides (N) were removed).

12. Sequences with low taxonomic precision (above phylum) were removed.

13. A random subset of 10 sequences was retained for each taxa.

14. For a given taxonomic rank, records with NA values were removed if they were not the only representative of the upper clade.

15. For Switzerland: Missing EPT species and IBCH families were searched in BOLD and merged with the reference database. The same filters and cleaning steps (2-14) were then applied.