-
Notifications
You must be signed in to change notification settings - Fork 4
Description
I noticed that when creating reference packages that have guaranteed sequences from TIGRFAM, the header gets truncated and as a result, when querying the NCBI, the sequences gets misclassified as "r__Root".
For example for NapA, here's the base treesapp create command
treesapp create -c NapA -p 0.85 --min_taxonomic_rank c -n 16 -i RefPkgs/Nitrogen_metabolism/Denitrification/NapA/ENOG501NS3T.faa --guarantee RefPkgs/Nitrogen_metabolism/Denitrification/NapA/TIGR01706.faa --cluster --trim_align --outdet_align --headless --fast --overwrite -o TS_Make_Lin_Table_For_Eval/Base/NapA/ --profile RefPkgs/Nitrogen_metabolism/Denitrification/NapA/TIGR01706.HMM --deduplicate --min_seq_length 600
For the TIGRFAM file, here are the sequence headers:
SP|Q56350|NAPA_PARDT/2-831
SP|P39185|NAPA_ALCEU/2-831
If you look at both the accession table and any trees that are generated for this package
These both get truncated to:
SP| r__Root
When they should be:
Q56350 r__Root; d__Bacteria; p__Pseudomonadota; c__Alphaproteobacteria; o__Rhodobacterales; f__Paracoccaceae; g__Paracoccus; s__Paracoccus pantotrophus
P39185 r__Root; d__Bacteria; p__Pseudomonadota; c__Betaproteobacteria; o__Burkholderiales; f__Burkholderiaceae; g__Cupriavidus; s__Cupriavidus necator
When removing the prefix of the headers, this fixes the issue, however, I'm wondering if the header truncation needs to be addressed.
However, this doesn't seem to be an issue when running treesapp with these prefixes in the base fasta input file (Example for RadA), or when treesapp update is used, after which the final clustered sequences go to treesapp create.
- TreeSAPP Version [e.g. 0.11.4]