Skip to content

Commit

Permalink
adding min hit treshold for krona output
Browse files Browse the repository at this point in the history
  • Loading branch information
maxibor committed Mar 22, 2018
1 parent 6b23408 commit 65130d8
Show file tree
Hide file tree
Showing 2 changed files with 19 additions and 7 deletions.
23 changes: 17 additions & 6 deletions bin/centrifuge2krona
Original file line number Diff line number Diff line change
Expand Up @@ -31,10 +31,16 @@ def get_args():
default=256,
help="Minimum centrifuge score to keep read. Default = 256")
parser.add_argument(
'-minhits',
default=5,
help="Minimum number of hits to include in krona chart"
)
parser.add_argument(
'-outpath',
default="./",
help="Output directory. Default = ./")


args = parser.parse_args()

infile = args.infile
Expand All @@ -43,10 +49,11 @@ def get_args():
taxo = args.tax
minlen = str(args.minlen)
minscore = str(args.minscore)
minhits = args.minhits
outpath = args.outpath


return(infile, index, taxo, lca, minlen, minscore, outpath)
return(infile, index, taxo, lca, minlen, minscore, minhits, outpath)

def get_basename(file_name):
if ("/") in file_name:
Expand All @@ -57,30 +64,34 @@ def get_basename(file_name):


if __name__ == "__main__":
infile, index, taxo, lca, minlen, minscore, outpath = get_args()
infile, index, taxo, lca, minlen, minscore, minhits, outpath = get_args()

basename = get_basename(infile)
if lca == True:
cmd = "centrifuge-kreport -x "+index+" --min-score "+minscore+" --min-length "+minlen+" "+infile+ " > "+basename+"_centriKraken_LCA_minScore_"+minscore+"_minLength_"+minlen+".out"
print(cmd)
os.system(cmd)
cmd = "awk '$2 > "+str(minhits)+"' "+basename+"_centriKraken_LCA_minScore_"+minscore+"_minLength_"+minlen+".out > "+basename+"_centriKraken_LCA_minScore_"+minscore+"_minLength_"+minlen+"_minhit"+str(minhits)+".out"
os.system(cmd)
if not taxo:
cmd = "ktImportTaxonomy -q 2 -t 5 -m 3-o "+basename+"_krona.html "+outpath+"/"+basename+"_centriKraken_LCA_minScore_"+minscore+"_minLength_"+minlen+".out"
cmd = "ktImportTaxonomy -q 2 -s 1 -t 5 -m 2 -o "+basename+"_krona.html "+outpath+"/"+basename+"_centriKraken_LCA_minScore_"+minscore+"_minLength_"+minlen+"_minhit"+str(minhits)+".out"
print(cmd)
os.system(cmd)
else:
cmd = "ktImportTaxonomy -q 2 -t 5 -m 3 -tax "+taxo+" -o "+basename+"_krona.html "+outpath+"/"+basename+"_centriKraken_LCA_minScore_"+minscore+"_minLength_"+minlen+".out"
cmd = "ktImportTaxonomy -q 2 -s 1 -t 5 -m 2 -tax "+taxo+" -o "+basename+"_krona.html "+outpath+"/"+basename+"_centriKraken_LCA_minScore_"+minscore+"_minLength_"+minlen+"_minhit"+str(minhits)+".out"
print(cmd)
os.system(cmd)
else:
cmd = "centrifuge-kreport -x "+index+" --no-lca --min-score "+minscore+" --min-length "+minlen+" "+infile+" > "+basename+"_centriKraken_noLCA_minScore_"+minscore+"_minLength_"+minlen+".out"
print(cmd)
os.system(cmd)
cmd = "awk '$2 > "+str(minhits)+"' "+basename+"_centriKraken_noLCA_minScore_"+minscore+"_minLength_"+minlen+".out > "+basename+"_centriKraken_noLCA_minScore_"+minscore+"_minLength_"+minlen+"_minhit"+str(minhits)+".out"
os.system(cmd)
if not taxo:
cmd = "ktImportTaxonomy -q 2 -t 5 -m 3 -o "+basename+"_krona.html "+outpath+"/"+basename+"_centriKraken_noLCA_minScore_"+minscore+"_minLength_"+minlen+".out"
cmd = "ktImportTaxonomy -q 2 -s 1 -t 5 -m 2 -o "+basename+"_krona.html "+outpath+"/"+basename+"_centriKraken_noLCA_minScore_"+minscore+"_minLength_"+minlen+"_minhit"+str(minhits)+".out"
print(cmd)
os.system(cmd)
else:
cmd = "ktImportTaxonomy -q 2 -t 5 -m 3 -tax "+taxo+" -o "+basename+"_krona.html "+outpath+"/"+basename+"_centriKraken_noLCA_minScore_"+minscore+"_minLength_"+minlen+".out"
cmd = "ktImportTaxonomy -q 2 -s 1 -t 5 -m 2 -tax "+taxo+" -o "+basename+"_krona.html "+outpath+"/"+basename+"_centriKraken_noLCA_minScore_"+minscore+"_minLength_"+minlen+"_minhit"+str(minhits)+".out"
print(cmd)
os.system(cmd)
3 changes: 2 additions & 1 deletion main.nf
Original file line number Diff line number Diff line change
Expand Up @@ -68,7 +68,7 @@ def helpMessage() {

//Pipeline version
version = "0.2.5"
version_date = "March 21th, 2018"
version_date = "March 22nd, 2018"

params.reads = "*_{1,2}.fastq.gz"
params.ctrl = "none"
Expand Down Expand Up @@ -812,6 +812,7 @@ if (params.aligner2 == "diamond"){

output:
set val(name), file("*.krona.html") into krona_res
set val(name), file("*_centriKraken_") into centri2kraken

script:
krona_out = name+".krona.html"
Expand Down

0 comments on commit 65130d8

Please sign in to comment.