You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
I'm trying to use the STAT tools to build a database from fasta sequences, and then using it to do metagenomics/taxonomic analyses.
So I'm following the tutorial in the build_db_and_run.sh script.
It says that we can do the identify_tax_ids part in multiple instances, but if we do, we have to use the tool called merge_kingdoms to combine results into a single file.
My problem is that there is no informations about the use of this tool.
The help of the tool is : need <tax.parents>
I don't understand what I should put in each argument (except for tax.parents).
Also, I'm using the default parameters :
KMER_LEN=32
DENSE_WINDOW=4 # 1 kmer of 4 for dense db (just for example)
SPARSE_WINDOW=128 # 1 kmer of 128 for sparse db (just for example)
But I don't know if I really should ?
Same question for MAX_KMER_DICTIONARY_SIZE=5000000 # This number should be roughly as max kmers expected * 2.
I don't really know how I could know the maximum number of kmers expected.
Thank you in advance !
The text was updated successfully, but these errors were encountered:
Hello,
I'm trying to use the STAT tools to build a database from fasta sequences, and then using it to do metagenomics/taxonomic analyses.
So I'm following the tutorial in the build_db_and_run.sh script.
It says that we can do the identify_tax_ids part in multiple instances, but if we do, we have to use the tool called merge_kingdoms to combine results into a single file.
My problem is that there is no informations about the use of this tool.
The help of the tool is : need <tax.parents>
I don't understand what I should put in each argument (except for tax.parents).
Also, I'm using the default parameters :
KMER_LEN=32
DENSE_WINDOW=4 # 1 kmer of 4 for dense db (just for example)
SPARSE_WINDOW=128 # 1 kmer of 128 for sparse db (just for example)
But I don't know if I really should ?
Same question for MAX_KMER_DICTIONARY_SIZE=5000000 # This number should be roughly as max kmers expected * 2.
I don't really know how I could know the maximum number of kmers expected.
Thank you in advance !
The text was updated successfully, but these errors were encountered: