Exception in denoising.py when using --denoising and --distance > 1 arguments

[gregdev00]

Hi I want to inform you, that the denoising.py tool throws exceptions if I use it with --denoising and --distance 3 arguments.

The issue appears to be caused by a typo in the variable name and missing functions that were available in previous versions.

Issue 1: Typo in Variable Name

The error occurs on line 1855 of denoising.py

It is: denoising_resized_seeds = tmpFiles.add( filename_woext + '_denoising_resizedSeeds.fasta' )

But it should be: denoising_resized_seeds = tmp_files.add( filename_woext + '_denoising_resizedSeeds.fasta' )

Issue 2: Missing Functions

Two functions that were previously available in clustering.py (before FROGS v5.0.0) are missing as well in denoising.py:

resizeSeed()
agregate_composition()

The script uses them, but they are not included in denoising.py anymore.

def resizeSeed(seed_in, seed_in_compo, seed_out):
    """
    @summary: add read abundance to seed sequence name
    @param seed_in : [str] Path to seed input fasta file
    @param seed_in_compo : [str] Path to seed input composition swarm file
    @param seed_out : [str] Path to seed output fasta file with abundance in name and sorted
    """
    dict_cluster_abond=dict()
    with open(seed_in_compo,"rt") as f:
        for idx,line in enumerate(f.readlines()):
            if not line.startswith("#"):
                cluster_name = "Cluster_" + str(idx+1) if not "FROGS_combined" in line.split()[0] else "Cluster_" + str(idx+1) + "_FROGS_combined"
                dict_cluster_abond[cluster_name]=sum([ int(n.split("_")[-1]) for n in line.strip().split()])
    f.close()

    FH_input = FastaIO( seed_in )
    FH_out=FastaIO(seed_out , "wt" )
    for record in FH_input:
        record.id += "_" + str(dict_cluster_abond[record.id])
        FH_out.write( record )
    FH_input.close()
    FH_out.close()

def agregate_composition(step1_compo , step2_compo, out_compo):
    """
    @summary: convert cluster composition in cluster in cluster composition in read (in case of two steps clustering)
    @param step1_compo : [str] Path to cluster1 composition in read (clustering step1)
    @param step2_compo : [str] Path to cluster2 composition in cluster1 (clustering step2) 
    @param out_composition : [str] Path to cluster2 composition in read
    """
    dict_cluster1_compo=dict()
    with open(step1_compo,"rt") as f:
        for idx,line in enumerate(f.readlines()):
            if "FROGS_combined" in line.split()[0]:
                dict_cluster1_compo["Cluster_"+str(idx+1)+"_FROGS_combined"]=line.strip()
            else:
                dict_cluster1_compo["Cluster_"+str(idx+1)]=line.strip()
    f.close()

    FH_out=open(out_compo,"wt")
    with open(step2_compo,"rt") as f:
        for line in f.readlines():
            compo=" ".join([dict_cluster1_compo["_".join(n.split('_')[0:-1])] for n in line.strip().split(" ")])
            FH_out.write(compo+"\n")

I hope you can fix this soon! Thanks for your hard work on FROGS.

[olivierrue]

Hi @gregdev00,
Thank you for reporting the problem. The bug has been fixed in version 5.0.1 which has just been released on the bioconda channel.