Separation of BOLD genus and species in SBDI export

[pragermh]

Description of the bug

Hi,

Related to #921 but much simpler. Here, the specificEpithet, infraspecificEpithet, and (for non-BIN matches) scientificName fields in the SBDI export from annotation against COIDB become misaligned when genus ends with an underscore followed by one or more 'X's.

Example:
genus: Malacostraca_XXX → specificEpithet: Malacostraca, infraspecificEpithet: XXXX

I can fix this with the code below, but I assume it should be quite straightforward to handle robustly in ampliseq parsing as well.

# Flag BOLD BINs
annotation[, isBIN := grepl("^BOLD:[A-Z0-9]+$", scientificName)]
# Fix mis-split names
annotation[grepl("_[X]+$", genus), `:=`(
    specificEpithet = "X",
    infraspecificEpithet = "",
    scientificName = ifelse(!isBIN, paste0(genus, specificEpithet), scientificName)
  ) 
]

Regards,
Maria

Command used and terminal output

Relevant files

No response

System information

No response

[erikrikarddaniel]

Can you give a couple of examples of input string (including in which rank they're now) and how they're supposed to be handled? I'm asking because "Malacostraca_XXX" doesn't look like it contains a specific epithet to me; it looks just like a genus name and I hence don't know how to interpret your example.

[pragermh]

In COIDB, Malacostraca_XXX is a genus value, and Malacostraca_XXXX is a species value, so there are no actual epithets, but in the SBDI export specificEpithet becomes 'Malacostraca' which is problematic.

[pragermh]

I'll send you some more examples when I get back to my office.

[pragermh]

Here is some output from annotation against COIDB 221216 (https://doi.org/10.17044/scilifelab.20514192.v2) using Ampliseq v2.12.0 VSEARCH:sintax. Note epithet and scientificName cols.

annotation-sub.tsv

I think the input from COIDB is in this file: bold_info_filtered.tsv, but the pattern is as described above, i.e. a 'genus' ending with '_X...' has a 'species' with an added 'X', and in the SBDI export all the X:s end up in the 'infraspecificEpithet' and the actual name part in the 'specificEpithet'.