A bunch of scripts to make QIIMists' life easier.
It is a simple script that makes QIIME-compatible id_to_tax maps starting from a fasta file containing the ids in the headers. Usage:
./id_to_tax_mapmaker.sh [fasta file] > id_to_tax.map
Here's the steps it goes through:
**1)**The script looks for the names.dmp and nodes.dmp files from the taxdump archive (ftp://ftp.ncbi.nlm.nih.gov/pub/taxonomy/taxdump.tar.gz) and for the big gi_taxid_nucl.dmp file (ftp://ftp.ncbi.nlm.nih.gov/pub/taxonomy/gi_taxid_nucl.dmp.gz); if not found in the same folder of the executable, it downloads them from NCBI.
2) The script cleans nodes.dmp and names.dmp of the non-scientific names that are essentially duplicates of the scientific ones, and could create confusion later on.
3) It reads the fasta file provided as argument, and matches the gi accession number (required) to the entire taxonomy from the db. The fasta header should be in the format:
>gi|XXXXXXX| Description...
or
>gi|XXXXXXX|ref|XXXXXXXX| Description...
The output will be a tab-separated table (as required by QIIME) with the identifier on the first column and the taxa levels on the second, e.g.:
gi|444303911|ref|NR_074334.1| Archaea;Euryarchaeota;Archaeoglobi;Archaeoglobales;Archaeoglobaceae;Archaeoglobus;Archaeoglobus fulgidus;Archaeoglobus fulgidus DSM 4304
The script has been tested on the NCBI's 16SMicrobial dataset (ftp://ftp.ncbi.nlm.nih.gov/blast/db/16SMicrobial.tar.gz) after converting it from blast_db format to pseudo-fasta format - we just need the gi IDs in the headers (blastdbcmd -db 16SMicrobial -outfmt '%f' -entry 'all' > 16SMicrobial.fasta). Warning: this works on blastdbcmd v.2.2.30. It stopped working with v.2.5.0, as the %f option does not output the GI in the header anymore.
This allows to run QIIME's assign_taxonomy.py using the following parameters:
--assignment_method blast
--id_to_taxonomy_fp id_to_tax.map
--reference_seqs_fp 16sMicrobial.fasta
obtaining characterization down to the species level.