file_suffix flag about gunc

[Biofarmer]

In addition, may I ask whether compressed fasta (.fna.gz) could be directly used by GUNC?
Thanks

[fullama]

If providing with --input_dir and --file_suffix .fna, I am wondering whether GUNC could make right action on those kind of genomes that contains .fna in the middle of names?

gunc will take everything before the first occurance of .fna in the file name and use that as the sample name in the output

may I ask whether --file_suffix .fna is still needed when --input_file is provide?

No, but the sample names in your output will contain the suffix, providing the suffix is only there to allow gunc to remove it from the input filename

may I ask whether compressed fasta (.fna.gz) could be directly used by GUNC?

Yes

any other questions let me know!

[Biofarmer]

Hi, many thanks for your reply. So, the function of --file_suffix is just to provide the file name in the output, there is no any effect on the detection of chimerism in genomes, do I understand correctly?

[fullama]

Correct! :)

[Biofarmer]

Okay, many thanks.

[Biofarmer]

Sorry, a further question about the database of progenomes or GTDB, may I ask which one is generally recommended to use for the detection of chimerism in genomes?
Thanks

[defleury]

Hi @Biofarmer !

Both databases work fine, we found little difference in accuracy. However, since the GUNC db based on proGenomes is smaller, it is faster to run so we use it by default.

[Biofarmer]

Okay, it is good to know. Thanks

[Biofarmer]

Hi, I am running GUNC for 10000 genomes with 5 threads, and it has been running 8 days, and now at Running Diamond period. There is no "diamond_output" folder in the output directory, is it normal?
In addition, may I ask whether it is possible to check the process of diamond? Because I want to know how long I still need to wait, if longer, I may kill the job and rerun it with more threads or split the genomes into parts.
PS, I have been run another 10000 genomes with 10 threads before, which was finished with 3 days.
Thanks

[fullama]

there is no way currently of seeing the progress of diamond, the run time can vary depending on the input.. but maybe you just want to run them in smaller batches? if you are running so many genomes at once i would increase both cpus and memory..

[Biofarmer]

Hi, thanks for your reply. In addition, if I understand correctly, the genecall files will be merged from all input genomes, if so, may I ask whether the label of each contig (text after ">" but before the first space, which is taken by prodigal as gene ID) should be unique for all input genomes? Or it does not matter? Thanks

[fullama]

it shouldnt matter.. they are merged but are tagged with the name of the genome file so they can be separated after diamond has run

[Biofarmer]

Okay...that's great. Thanks. The merged genecall files is intermediate, and has been deleted once finished and cannot be seen, right?

[Biofarmer]

In addition, as the --temp_dir directory by default is Current working directory. If I submit several jobs in the same working directory at once with different output directory, may I ask whether GUNC will select the right temporary files from the same working directory? Is the temporary file of each job with different names? Thanks

[Biofarmer]

Hi, may I ask the answers from questions as above?

[fullama]

I think it would be fine but try it out and see to be sure..

[Biofarmer]

Hi, many thanks for your confirmation.

[Biofarmer]

Hi, I just used GUNC to check the genome from NCBI, GCF_902703415.1_Combinated_assembly_ONT_-_Illumina_genomic.fna. It failed, however, when changed the genome name without _-_ to GCF_902703415.1_Combinated_assembly_ONT_---_Illumina_genomic.fna, or GCF_902703415.1_Combinated_assembly_ONT_--_Illumina_genomic.fna, or GCF_902703415.1_Combinated_assembly_ONT__Illumina_genomic.fna, or
GCF_902703415.1_Combinated_assembly_ONT-Illumina_genomic.fna, or
GCF_902703415.1_Combinated_assembly_ONT_Illumina_genomic.fna. Anyone worked with GUNC. May I ask why only genome name containing _-_ format did not work with GUNC?

And GUNC worked for genome GCF_902109435.1_40087_F01_genomic.fna, but there was no value in output, is it due to its small size of genome?

Thanks

[Biofarmer]

Another question: there is a slight difference in the number of n_genes_mapped in the output when running one genome individually (--input_fasta genome.fna) or running a few genomes together (--input_dir genome_folder/ --file_suffix .fna). May I ask whether it is normal and why?
Thanks

[fullama]

samples with _-_ dont work because internally gunc uses it as a delimiter to label sequences with the samples name when merging them together.. I didnt think anyone would ever use _-_ in a sample name.. ill see if i can change it for the next version..

[fullama]

Another question: there is a slight difference in the number of n_genes_mapped in the output when running one genome individually (--input_fasta genome.fna) or running a few genomes together (--input_dir genome_folder/ --file_suffix .fna). May I ask whether it is normal and why? Thanks

Can you give an example of where the output differs, so i can look more closely?

[Biofarmer]

samples with _-_ dont work because internally gunc uses it as a delimiter to label sequences with the samples name when merging them together.. I didnt think anyone would ever use _-_ in a sample name.. ill see if i can change it for the next version..

Thanks for reply.

[Biofarmer]

Another question: there is a slight difference in the number of n_genes_mapped in the output when running one genome individually (--input_fasta genome.fna) or running a few genomes together (--input_dir genome_folder/ --file_suffix .fna). May I ask whether it is normal and why? Thanks

Can you give an example of where the output differs, so i can look more closely?

Hi, just take the genomes from NCBI (GCF_902703415.1 and GCF_900232175.1) for example, the n_genes_mapped is 5145 and 6693 when tested together, and the value will be 5157 and 6740 respectively when tested individually. The overall conclusion is almost same.
By the way, may I ask whether the result of pass.GUNC is based on the third digit of clade_separation_score after decimal? Because I see same genomes that are reported with clade_separation_score 0.45, sometimes are given False or sometimes True for pass.GUNC. So I am wondering the clade_separation_score is just reported with two digits after decimal, but judgement for pass.GUNC is based on at least three digits after decimal.

[fullama]

Hi, just take the genomes from NCBI (GCF_902703415.1 and GCF_900232175.1) for example, the n_genes_mapped is 5145 and 6693 when tested together, and the value will be 5157 and 6740 respectively when tested individually. The overall conclusion is almost same.

this will be fixed in the next version of gunc

may I ask whether the result of pass.GUNC is based on the third digit of clade_separation_score after decimal?

yes, the output will be amended to include more decimal places in the next version also