TODO list

[shenwei356]

• Search: supporting specifying TaxID, e.g., only searching from a species or a genus.
  • No need to rebuild the index. Just filter matches in the seed-matching step .
  • Existing information:
    • genomes.map.bin stores genome id - internal id pairs.
  • File needed:
    • a genome accession -> taxid mapping file
    • taxdump files from NCBI or created by TaxonKit
  • Relationship: internal id -> taxid
  • Check: isAChild = LCA(taxid_target, taxid_test) == taxid_target.
  • implemented: https://github.com/shenwei356/LexicMap/tree/search-by-taxid
• Create a table to explain the changes and compatibility of the index format.
• Add a daemon process for searching via RESTful API.
• Add a utility tool to edit genome names in the index via a regular expression, which only needs to edit the file genomes.map.bin.
• utils subseq: accept search result as input, for batch sequence extraction.
  • parallelise it.
• add a new command to combine multiple indexes (dozens)

[ypchan]

Hi Dr. Shen,

Are there any plans to support LCA- or majority-vote-based taxonomic assignment using LexicMap results?

For metagenomic data, I have tried assigning contigs to viruses, bacteria, archaea, fungi, and protozoa using the NR database, but many contigs could not be assigned to any group. In contrast, LexicMap can use both protein-coding and non-coding regions to identify related reference sequences, which could be very helpful for taxonomic assignment.

Yanpeng

[shenwei356]

Are there any plans to support LCA- or majority-vote-based taxonomic assignment using LexicMap results?

Sorry, it's another project.

But you can write some code to parse the search results. This simple demo might be useful.