Thanks for this nice tool! Is it possible to add advice in the readme text on how produce coverage.txt and mapping.sam ?

Hi,
Do you have any plan to release a Docker/Singularity version of COBRA?
Thanks

ImportError Traceback (most recent call last)
Cell In[23], line 4
2 path='show'
3 test_path=("data1/densiflorus (plastid)- MT740250.1 retained.fasta")
----> 4 BCAWT_auto_test.auto_test(path,test_path)
5 BCAWT_auto_test.auto_check_file(path)

File ~/miniconda3/lib/python3.11/site-packages/BCAWT/BCAWT_auto_test.py:24, in auto_test(path, test_file)
21 from BCAWT import BCAWT
22 file = open(test_file, "r")
---> 24 BCAWT.BCAW(main_fasta_file = [file] , save_path= path, Auto=True)

File ~/miniconda3/lib/python3.11/site-packages/BCAWT/BCAWT.py:34, in BCAW(main_fasta_file, save_path, ref_fasta_file, genetic_code_, Auto)
32 #from Bio.Alphabet import IUPAC
33 from Bio.Seq import Seq
---> 34 from Bio.SeqUtils import GC
35 from Bio.Data import CodonTable
36 #from Bio.Alphabet import generic_dna

ImportError: cannot import name 'GC' from 'Bio.SeqUtils' (/home/u1/miniconda3/lib/python3.11/site-packages/Bio/SeqUtils/init.py)

Hi there,

I'm excited about the potential this tool offers!

Currently, I'm receiving an error in the python command about the BLAST db COBRA needs:

$ python cobra.py 
-f final.contigs.renamed.fa /
-q samp_471_concoct_280.fa /
-c coverage_tail.txt /
-m final.contigs.renamed.sam /
-a megahit /
-mink 21 /
-maxk 99

Command line error:

Building a new DB, current time: 06/30/2023 13:24:18
New DB name:  samp_471_concoct_280.fa_COBRA/blastdb_1.fa
New DB title:  blastdb_1.fa
Sequence type: Nucleotide
Keep MBits: T
Maximum file size: 1000000000B
BLAST options error: File blastdb_1.fa is empty
BLAST Database error: No alias or index file found for nucleotide database [blastdb_1.fa] in search path [samp_471_concoct_280.fa_COBRA::/reference-data/blast:]

Any guidance would be greatly appreciated.

Hi Dr Chen
After conda activate cobra, we tried to run cobra pipeline, with this error occured as
"cobra-meta" missing
do i miss to install something?

thanks.

Hey LinXing,

First, thanks for developing this super useful tool! I've so far ran it on a couple hundred virome samples and it's working great! I have stumbled across an issue with 2 samples though that I was wondering if you could help me with. For each of the 2 metagenomes I get this error (but obviously a different key value differing the messages):

Traceback (most recent call last):                                                                                                                                            
  File ".snakemake/conda/e7dfc9b82d378646d96b58efd8350561_/bin/cobra-meta", line 12, in <module>                  
    sys.exit(main())                                                                                                                                                          
  File ".snakemake/conda/e7dfc9b82d378646d96b58efd8350561_/lib/python3.10/site-packages/cobra.py", line 1598, in m
ain                                                                                                                                                                           
    extended_circular_fasta.write(header2joined_seq[contig2extended_status[contig]] + '\n')                                                                                   
KeyError: 'NODE_404_length_6556_cov_28.183972'

I checked for these keys in the whole contig set (the scaffold.fasta output by metaspades); query file (all scaffolds >2.5kbp from the metaspades scaffolds.fasta); and the coverage file. It is present in each of those files with no additional characters. So, I'm a bit confused with what the source of the problem could be (especially since the couple of hundred other samples i've processed worked fine!). Any help you could give would be much appreciated!

The the log file looks like this:

1. INPUT INFORMATION
# Assembler: metaSPAdes
# Min-kmer: 21
# Max-kmer: 55
# Overlap length: 55 bp
# Read mapping max mismatches for contig linkage: 2
# Query contigs: /data/san/data0/users/chris/data/processed/assemblies/spades_CTRLM08LongB03Faeces/scaffolds_over_2.5kb.fasta
# Whole contig set: /data/san/data0/users/chris/data/processed/assemblies/spades_CTRLM08LongB03Faeces/scaffolds.fasta
# Mapping file: /data/san/data0/users/chris/data/processed/assemblies/spades_CTRLM08LongB03Faeces/scaffolds_sorted.bam
# Coverage file: /data/san/data0/users/chris/data/processed/assemblies/spades_CTRLM08LongB03Faeces/scaffolds.coverage.formatted.txt
# Output folder: /data/san/data0/users/chris/data/processed/assemblies/spades_CTRLM08LongB03Faeces/cobra

2. PROCESSING STEPS
[01/23] [2024/02/16 00:22:53] Reading contigs and getting the contig end sequences. A total of 255643 contigs were imported.
[02/23] [2024/02/16 00:22:58] Getting shared contig ends.
[03/23] [2024/02/16 00:23:00] Writing contig end joining pairs.
[04/23] [2024/02/16 00:23:00] Getting contig coverage information.
[05/23] [2024/02/16 00:23:00] Getting query contig list. A total of 2314 query contigs were imported.
[06/23] [2024/02/16 00:23:00] Getting contig linkage based on sam/bam. Be patient, this may take long.
[07/23] [2024/02/16 00:23:38] Parsing the linkage information.
[08/23] [2024/02/16 00:23:43] Detecting self_circular contigs.                                                                                                                     "zEQ$" 23:44 15-Feb-24
[09/23] [2024/02/16 00:23:43] Detecting joins of contigs. 10%, 20%, 30%, 40%, 50%, 60%, 70%, 80%, 90%, 100% finished.
[10/23] [2024/02/16 00:23:46] Saving potential joining paths.
[11/23] [2024/02/16 00:23:46] Checking for invalid joining: sharing queries.
[12/23] [2024/02/16 00:23:47] Getting initial joining status of each query contig.
[13/23] [2024/02/16 00:23:47] Getting final joining status of each query contig.
[14/23] [2024/02/16 00:23:47] Getting the joining order of contigs.
[15/23] [2024/02/16 00:23:47] Getting retrieved contigs.
[16/23] [2024/02/16 00:23:47] Saving joined seqeuences.
[17/23] [2024/02/16 00:23:47] Checking for invalid joining using BLASTn: close strains.
[18/23] [2024/02/16 00:23:48] Saving unique sequences of "Extended_circular" and "Extended_partial" for joining checking.

Dear Linxing,

Thank you for providing an excellent tool for viral metagenomic studies. I have a question about the usage of COBRA:

I want to improve the genome quality of viruses recovered from multi-sample coassembled metagenomes (such as A1, A2, A3). In this case, I merged the mapping files of each metagenome (i.e., A1.bam A2.bam A3.bam > merge.bam) as the bam input and calculate the coverage file using jgi_summarize_bam_contig_depths (the first column was contig name, the second to fourth columns were the coverage of A1, A2, A3) as the coverage input. I'm not sure if my inputs to COBRA were correct.

Regards,
Mengzhi

Dear developers. Could you clarify please. Why is the group Extended_partial based on checkV results circularized (with DTR)? Can I take these sequences as putative complete genomes?

I read the paper. In discussion you mentioned that "COBRA can also be applied to microbial genomes and whole metagenonmes" and "using COBRA only to extend the subset of longer contigs". How about to use COBRA as a binning refiner? Although some refiner has been reported, such as https://github.com/dparks1134/RefineM and https://apcamargo.github.io/magpurify2, purifing bins is difficult in extremely complex microbial community. Maybe COBRA have more performance as a binning refiner.

Hello, thank you for the great tool!
My question is that can the assembly come from several different but similar sources? I have gut microbiome samples from the same organisms but from different sampling times and indivduals and made a coassembly from them.
Thanks, Balazs

Hi @linxingchen

I've noticed that sometimes COBRA prints contig names and numbers on the screen without a clear explanation of what they represent. It appears that the code in these lines is responsible for this. Could it be leftover debug code?

Thanks for this great tool! However, I have a few questions on input for cobra

1. Now I'm using megahit to assembly (I just know from #3 that it may generate many chimeric contigs, but spades may run out of memory when assembly environmental samples). In megahit intermediate results, it will provide a file named ./intermediate_contigs/k141.contigs.fa, in which there are many small contigs < 200 bp (which will be filtered in ./final.contigs.fa). My question is, which contig file is more recommended to be used as --fasta FASTA input?
2. Is it possible to just use the final.contigs.fa as --query QUERY file, or a filtered version with all contigs longer than given length (i.e. 1000 or 2500 bp)? In another words, can cobra be used before virus contigs annotation and MAG binning?

Regards,
hwrn

Hi,

You mention that COBRA has so far only been tested on assembled contigs from metaSPAdes, IDBA_UD, and MEGAHIT. Would you recommend to use it on viral contigs that have been assembled using single cell mode (--sc) from SPAdes (used in our study for metaviromes)?

Thank you,
Asier

Dear Linxing,

Good afternoon. Thank you for this excellent paper and tool! I read this paper and shared it with my lab members. I have some questions about the application of COBRA:

1. For the metaSPAde you mentioned in the README file and paper, have you tested the MetaviralSPAdes(https://academic.oup.com/bioinformatics/article-abstract/36/14/4126/5837667) which is the specific viral mode for assembly? I'm working on some metagenomic data and trying to find the viral sequence in it. I think it would be a little different if we use metaSPAde and then identify viral contigs by other tools like VIBRANT or directly identify viral contigs based on MetaviralSPAdes.
   I think it could be a similar question with closed#22.

2. We're working on the Mitochondria sequence also, and do you think COBRA could be used into others like mitochondria sequences, to re-assemble those sequences, using reference genome sequence as input1 and assembled results as input2? Or we have to use those inputs from the same data?

Thank you for your time and patience!

Best Regards,
Xinpeng

Hi Linxing:
Thank you for developing this nice tool! I am curious about what's the functions does COBRA have in the classic virome pipeline as it's a new software.
For example, I used MEGAHIT to get contigs from metagenome reads, and then I used geNomad to identified the viral contigs from the total contigs, and then if I use COBRA in the follow step (i.e. put the geNomad results:<prefix>_summary/<prefix>_virus.fna as the input of COBRA) , does COBRA help me to bin the identified viral contigs together to get a higher completeness here? Can it work before or after geNomad well? I read about that COBRA can identify more circular viral genome and huge phage. Can I consider COBRA as a binner tool or a circular/huge phages identifier?
Thanks!

Hi Dr Chen：
I try to run COBRA today with

source activate cobra
cobra-meta -f Megahit_out/611PES1/611PES1_megahit.contigs.fa -q checkv_in/611RW.fa -o COBRA/611_RW_MV_COBRA -c coverage/611RW_coverage.txt -m mapping/611RW-MV.sam -a megahit -mink 21 -maxk 141
conda deactivate

and it works at beginning.
But it stop at cobra.py, line 1531:r = open('blastdb_2.vs.blastdb_1', 'r')
The error is

makeblastdb: error while loading shared libraries: libzstd.so.1: cannot open shared object file: No such file or directory
blastn: error while loading shared libraries: libzstd.so.1: cannot open shared object file: No such file or directory
Traceback (most recent call last):
  File "/data01nfs/user/songq/anaconda3/envs/cobra/bin/cobra-meta", line 10, in <module>
    sys.exit(main())
  File "/data01nfs/user/songq/anaconda3/envs/cobra/lib/python3.8/site-packages/cobra.py", line 1531, in main
    r = open('blastdb_2.vs.blastdb_1', 'r')
FileNotFoundError: [Errno 2] No such file or directory: 'blastdb_2.vs.blastdb_1'

blastdb_1.fa and blastdb_2.fa are written and not empty, and I also install BLAST in cobra env, so I guess The error was not caused by BLAST.
Please help me solve this problem, I am willing to provide you with any other information you need.
Thanks

Cobra is an interesting software, but I have some questions. Is the input data must be viral metagenome data? I'm not sure if cobra could works for common metagenome data and will cobra produce some bacteria or archaea MAGs? Thanks.

Hi Dr. Chen,
I encountered this problem while testing COBRA：
My command:

cobra-meta -f MG_CE.fasta -q MG_CE_query_id.txt -o cobra-test-out -c coverage.txt -m MG_CE-cobra-fs.bam -a megahit -mink 21 -maxk 141 -mm 2

Error:

usage: cobra-meta [-h] -q QUERY -f FASTA -a {idba,megahit,metaspades} -mink MINK -maxk MAXK -m MAPPING -c COVERAGE [-lm LINKAGE_MISMATCH] [-o OUTPUT] [-t THREADS] [-v]
cobra-meta: error: unrecognized arguments: 2

And it can run after deleting -mm 2.
Or I may input -lm 2 instand of -mm 2? If that's the case, you may need to modify the How to run section in the README.md.

Thanks!

Hi Linxing,

Great tool. I'm really interested in using it with some of my datasets. One question that came up in discussing COBRA with others was how it handles cases where e.g. Query Contigs Q1 and Q2 could be joined by non-query contigs, e.g. Q1-c1-c2-Q2. Apologies if I missed this part of the manuscript but does COBRA have the ability to handle this, or does it start from the assumption that it is only looking for ways it can extend query contigs in isolation?

Best,

Luke

Hi Linxing! I just wanted to recommend adding detailed installation instructions for a conda environment that users could just copy and paste. I know it's silly, but many users are not savvy in python but have conda available, and I think it would extend your user list if you just do this one small thing.

Hi, @linxingchen
According to issues #2, #22, and #27, I think maybe I can extend the virus contigs with COBRA then binning the new contigs with binner tools like vRhyme to get virus genomes?

Hi Xinglin,

Thank you for your outstanding paper and tool!

I'm looking to identify RNA viruses in the metatranscriptome. After using Megahit for assembly, I'm curious if COBRA could be used to improve the quality of contigs.

Best,
Shujie

Hi.

After your suggestion last time, I reassembled my data with metaSpades (single assembly) using (-k 21,33,55,77,99,127).

From the output assembly, I filtered viral contigs using Vibrant (1983 putative phages were identified).

From here I fed these predicted phages into vRhyme and COBRA

This is the result of vRhyme for circular contigs.

Scaffold	type	mismatches	length	repeat
NODE_31851_length_2369_cov_1.518733	DTR	0	127	CCTCCATCAAGATGAATACAACTCCATTTACCAAGATTGTTGGCAAAATAGCTGTCATCCTCGGCAAAATTGAAACTGTCATGCACAAAGACAGCTACGCGCGTCACAGTAATGCGCCAGCGCCCAT
NODE_18158_length_3813_cov_11.723820	DTR	0	127	TAGCCGGTTGCTCTGGCACCTCTCGTACCGGCTCCTGCTTCACCCGCAGAGGGTTGATGTCCCGGCCCGAAAGGATGTAGTCCCCCGCGCCGAATCGCCAGCGGAAGGTAACCGCCAGTTCCGCCAC
NODE_14185_length_4746_cov_1.590171	DTR	0	127	ACTTCTGTTTAAATTCTTCTTTGTTAATATTTGTGTACCAGCCCCATCCATTGTAATCATGTCCGCCTGTCAAATGGTATGCTCTTTTTAATGAAGTATAATTCCCTGAAAGATGGACGCCTGTAAT
NODE_421_length_73347_cov_5.589907	ITR	0	116	GGATTGCCACCAGATATGACAAATTGGATACGTCTTTTCTCGCCTTCGTTTATATTGCCGCTATCGCTATTTTGTTAATTTAGCTCATCCCTCCTTATTTTTCAAACAACCCCTAA
NODE_4030_length_14370_cov_3.630134	DTR	0	40	ATAGAAAATAGAATAAAAAACTCATTTGTCACATGAAAAA
NODE_18745_length_3708_cov_3.479754	ITR	1	111	AAATATTTCCGATATCTATCAAAACAAAAGCGGTGGCGAATCAATCAAAACGAAGCGTTTCATACAGCAATTTTACCAGGGCTTTTGTCTTAAGTTCGCGCCTATGGGGCT

COBRA identified only 2 circular contigs

SeqID	Length	Coverage	GC	Ns	DTR_length	
NODE_31851_length_2369_cov_1.518733_self_circular	2369	10.331	57.957	0	127	
NODE_18158_length_3813_cov_11.723820_self_circular	3813	77.942	57.383	0	127	

COBRA even extended one circular contig as "COBRA_category_ii-b_extended_partial_unique"

SeqID	Length	Coverage	GC	Ns	
NODE_421_length_73347_cov_5.589907_extended_partial	88998	38.204	57.323	0	

Rest are "COBRA_category_ii-c_extended_failed"

SeqID	Length	Coverage	GC	Ns
NODE_4030_length_14370_cov_3.630134	14370	24.458	53.598	0	
NODE_18745_length_3708_cov_3.479754	3708	22.521	53.479	0

and "COBRA_category_iii_orphan_end"

SeqID	Length	Coverage	GC	Ns
NODE_14185_length_4746_cov_1.590171	4746	10.854	35.04	0

So, here are my questions

1. Is it possible for COBRA to identify all DTRs and ITRs?
2. As COBRA extended one of the vRhyme predicted circular contig, how should I interpret COBRA result? Is it correct to be extended by COBRA?
3. VIBRANT output is sometimes fragmented contigs whose name is not present in my assembly. These names are not recognised by COBRA. Is there any way for these fragments to be included? (In this sample there are 73 query)

Query NODE_464_length_69899_cov_2.489623_fragment_1 is not in your whole contig fasta file, please check!

Hello,
I am running into this error:

cobra-meta -f final.contigs.fa -q test.fa -c coverage.txt -m sorted_output.sam -a idba -mink 20 -maxk 140

/home/pck/anaconda3/envs/cobra/lib/python3.8/site-packages/Bio/SeqUtils/init.py:144: BiopythonDeprecationWarning: GC is deprecated; please use gc_fraction instead.
warnings.warn(
Traceback (most recent call last):
File "/home/pck/anaconda3/envs/cobra/bin/cobra-meta", line 10, in
sys.exit(main())
File "/home/pck/anaconda3/envs/cobra/lib/python3.8/site-packages/cobra.py", line 878, in main
if header2len[line.reference_name] > 1000:
KeyError: 'k141_2210 flag=1 multi=2.0000 len=345'

k141 is the very first contig in my input file, the only output file besides the log is COBRA_end_joining_pairs.txt

log:

1. INPUT INFORMATION

Assembler: IDBA_UD

Min-kmer: 20

Max-kmer: 140

Overlap length: 139 bp

Read mapping max mismatches for contig linkage: 2

Query contigs: /home/pck/metagenome_seq/megahit_output/test.fa

Whole contig set: /home/pck/metagenome_seq/megahit_output/final.contigs.fa

Mapping file: /home/pck/metagenome_seq/megahit_output/sorted_output.sam

Coverage file: /home/pck/metagenome_seq/megahit_output/coverage.txt

Output folder: /home/pck/metagenome_seq/megahit_output/test.fa_COBRA

2. PROCESSING STEPS
   [01/23] [2024/02/09 13:12:51] Reading contigs and getting the contig end sequences. A total of 3468 contigs were imported.
   [02/23] [2024/02/09 13:12:51] Getting shared contig ends.
   [03/23] [2024/02/09 13:12:51] Writing contig end joining pairs.
   [04/23] [2024/02/09 13:12:51] Getting contig coverage information.
   [05/23] [2024/02/09 13:12:51] Getting query contig list. A total of 1 query contigs were imported.
   [06/23] [2024/02/09 13:12:51] Getting contig linkage based on sam/bam. Be patient, this may take long.

Thank you for this amazing tool!

I installed cobra using micromamba. biopython=1.81 is installed but the GC deprecation issue still ensued.

$ cobra-meta -f pelv_comb_mghit_fixed.fa -q bin.10.fa -c pelv.comb.cov.txt -m sorted_pelv_comb.bam -a megahit -mink 21 -maxk 141 -t 8

/Users/andriangajigan/mambaforge/envs/cobra/lib/python3.8/site-packages/Bio/SeqUtils/init.py:144: BiopythonDeprecationWarning: GC is deprecated; please use gc_fraction instead.
warnings.warn(
Traceback (most recent call last):
File "/Users/andriangajigan/mambaforge/envs/cobra/bin/cobra-meta", line 10, in
sys.exit(main())
File "/Users/andriangajigan/mambaforge/envs/cobra/lib/python3.8/site-packages/cobra.py", line 818, in main
cov[line[0]] = round(float(line[1]), 3)
ValueError: could not convert string to float: 'contigLen

$ conda list
packages in environment at /Users/andriangajigan/mambaforge/envs/cobra:

Name Version Build Channel
biopython 1.81 py38hae2e43d_1 conda-forge
blast 2.15.0 pl5321h91c44f7_1 bioconda
bzip2 1.0.8 h10d778d_5 conda-forge
c-ares 1.26.0 h10d778d_0 conda-forge
ca-certificates 2024.2.2 h8857fd0_0 conda-forge
cobra-meta 1.2.2 pyhdfd78af_0 bioconda
curl 8.5.0 h726d00d_0 conda-forge
entrez-direct 16.2 h193322a_1 bioconda
gettext 0.21.1 h8a4c099_0 conda-forge
krb5 1.21.2 hb884880_0 conda-forge
libblas 3.9.0 21_osx64_openblas conda-forge
libcblas 3.9.0 21_osx64_openblas conda-forge
libcurl 8.5.0 h726d00d_0 conda-forge
libcxx 16.0.6 hd57cbcb_0 conda-forge
libdeflate 1.18 hac1461d_0 conda-forge
libedit 3.1.20191231 h0678c8f_2 conda-forge
libev 4.33 h10d778d_2 conda-forge
libffi 3.4.2 h0d85af4_5 conda-forge
libgfortran 5.0.0 13_2_0_h97931a8_2 conda-forge
libgfortran5 13.2.0 h2873a65_2 conda-forge
libiconv 1.17 hd75f5a5_2 conda-forge
libidn2 2.3.7 h10d778d_0 conda-forge
liblapack 3.9.0 21_osx64_openblas conda-forge
libnghttp2 1.58.0 h64cf6d3_1 conda-forge
libopenblas 0.3.26 openmp_hfef2a42_0 conda-forge
libsqlite 3.44.2 h92b6c6a_0 conda-forge
libssh2 1.11.0 hd019ec5_0 conda-forge
libunistring 0.9.10 h0d85af4_0 conda-forge
libzlib 1.2.13 h8a1eda9_5 conda-forge
llvm-openmp 17.0.6 hb6ac08f_0 conda-forge
ncbi-vdb 3.0.0 pl5321h9722bc1_0 bioconda
ncurses 6.4 h93d8f39_2 conda-forge
numpy 1.24.4 py38h9a4a08f_0 conda-forge
openssl 3.2.1 hd75f5a5_0 conda-forge
pcre 8.45 he49afe7_0 conda-forge
perl 5.32.1 7_h10d778d_perl5 conda-forge
perl-archive-tar 2.40 pl5321hdfd78af_0 bioconda
perl-carp 1.50 pl5321hd8ed1ab_0 conda-forge
perl-common-sense 3.75 pl5321hd8ed1ab_0 conda-forge
perl-compress-raw-bzip2 2.201 pl5321h775f41a_0 conda-forge
perl-compress-raw-zlib 2.202 pl5321h775f41a_0 conda-forge
perl-encode 3.19 pl5321hb7f2c08_0 conda-forge
perl-exporter 5.74 pl5321hd8ed1ab_0 conda-forge
perl-exporter-tiny 1.002002 pl5321hd8ed1ab_0 conda-forge
perl-extutils-makemaker 7.70 pl5321hd8ed1ab_0 conda-forge
perl-io-compress 2.201 pl5321h7133b54_2 bioconda
perl-io-zlib 1.14 pl5321hdfd78af_0 bioconda
perl-json 4.10 pl5321hdfd78af_0 bioconda
perl-json-xs 2.34 pl5321hcd10b59_5 bioconda
perl-list-moreutils 0.430 pl5321hdfd78af_0 bioconda

Thanks in advance!

Hello,

I have noticed that with some COBRA runs, COBRA randomly aborts after "Saving joining summary of "Extended_circular" and "Extended_partial" query contigs". The log files just end there, and the fasta files in question are missing. Everything else is there. Instead of the missing self circular files, I get what appears to be blast databases (blastdb_1 and blastdb_2) as output.

Rerunning the metagenomic analysis on the exact same original fastq files, starting with megahit assembly, it randomly works fine. All output files up until running coverage_transfer.py are identical (with slight contig naming differences because of megahit), I can not tell a difference in runs up until COBRA comes in the picture, and even log and debug files are identical up until the "Saving joining summary of "Extended_circular" and "Extended_partial" query contigs"

Very strange bug, and not sure how to figure out what's wrong since it appears random. Happy to provide successful/unsuccessful run data, just not sure what would help and don't want to dump it all here.

Hi.

Thank you for making this amazing tool.

I am currently using this tool to improve my metagenomic viral contigs obtained from MegaHIT.

So, I have 80 shotgun metagenomic samples which I did coassembly in MegaHIT followed by viral contig detection by VirSorter2.

I want to extend my viral contigs by COBRA, but it not showing any progress for the past 6 days.

This is the code I use (I am using my institute cluster)

python cobra.py -f ~/final.contigs.fa -q ~/final.contigs.phages_combined.fna \
-o VirSorter -c ~/CoverM/For_virus/Coverage.txt -m ~/MegaHit_final.contigs.fa.All.sorted.bam \
-a megahit -mink 21 -maxk 77

This is the logfile

1. INPUT INFORMATION
# Assembler: MEGAHIT
# Min-kmer: 21
# Max-kmer: 77
# Overlap length: 77 bp
# Read mapping max mismatches for contig linkage: 2
# Query contigs: /lustre7/home/bhimbiswa/MAGs/Virus/COBRA/final-viral-combined.fa
# Whole contig set: /lustre7/home/bhimbiswa/MAGs/MegaHit/output_folder/final.contigs.fa
# Mapping file: /lustre7/home/bhimbiswa/MAGs/CoverM/For_virus/bam_files/MegaHit_final.contigs.fa.All.sorted.bam
# Coverage file: /lustre7/home/bhimbiswa/MAGs/CoverM/For_virus/Coverage.txt
# Output folder: /lustre7/home/bhimbiswa/MAGs/Virus/COBRA/Virsorter2

2. PROCESSING STEPS
[01/22] [2023/06/07 14:43:28] Reading contigs and getting the contig end sequences. A total of 1259677 contigs were imported.
[02/22] [2023/06/07 14:44:13] Getting shared contig ends.
[03/22] [2023/06/07 14:44:34] Writing contig end joining pairs.
[04/22] [2023/06/07 14:44:35] Getting contig coverage information.
[05/22] [2023/06/07 14:44:39] Getting query contig list. A total of 47249 query contigs were imported.
[06/22] [2023/06/07 14:44:40] Getting contig linkage based on sam/bam. Be patient, this may take long.
[07/22] [2023/06/08 08:29:30] Detecting self_circular contigs. 
[08/22] [2023/06/08 09:14:19] Detecting joins of contigs. 

I can understand that I have lots of query files that why this is very slow, but is there any way I can make this faster?

P.S.: COBRA worked fine with Phigaro viral contigs (only 285 query contigs).

On a new installation of cobra I run into this error:

❯ cobra-meta -h                                                                                  (cobra)
Traceback (most recent call last):
  File "/lila/home/funnellt/mambaforge/envs/cobra/bin/cobra-meta", line 6, in <module>
    from cobra import main
  File "/lila/home/funnellt/mambaforge/envs/cobra/lib/python3.8/site-packages/cobra.py", line 13, in <module>
    from Bio.SeqUtils import GC
ImportError: cannot import name 'GC' from 'Bio.SeqUtils' (/lila/home/funnellt/mambaforge/envs/cobra/lib/python3.8/site-packages/Bio/SeqUtils/__init__.py)

I installed cobra using the instructions in the readme, which installed Biopython 1.82

❯ python                                                                                         (cobra)
Python 3.8.18 | packaged by conda-forge | (default, Dec 23 2023, 17:21:28)
[GCC 12.3.0] on linux
Type "help", "copyright", "credits" or "license" for more information.
>>> import Bio
>>> Bio.__version__
'1.82'

Looking through the Biopython NEWS.rst file, it looks like the GC module has been deprecated
https://github.com/biopython/biopython/blob/master/NEWS.rst

Thank you very much for your software output, which provides a new workflow for viral group research!

Building a new DB, current time: 03/09/2024 10:19:13
New DB name:   /home/zhongpei/hard_disk_sda2/zhongpei/Virome/rawdata/upload_20230812/zhongpei_analyse/fastp/final_assembly/clean_TPM_Decontam_contig/Vir_result/AF11_metaSPAdes_COBRA/blastdb_1.fa
New DB title:  blastdb_1.fa
Sequence type: Nucleotide
Keep MBits: T
Maximum file size: 3000000000B
BLAST options error: File blastdb_1.fa is empty
BLAST Database error: No alias or index file found for nucleotide database [blastdb_1.fa] in search path [/home/zhongpei/hard_disk_sda2/zhongpei/Virome/rawdata/upload_20230812/zhongpei_analyse/fastp/final_assembly/clean_TPM_Decontam_contig/Vir_result/AF11_metaSPAdes_COBRA::]

When I run COBRA, echo gives the following error
But it seems my process is complete. Is this reasonable? And is it a "BLAST error" because my -q sequence is not extended at all?
here is my log file
log.txt
It would be extremely helpful to hear from you!

Hello,

First of all, thank you for developing this great tool!

I recently read your published paper.
I would like to ask about the current recommendations or precautions for applying COBRA to stool whole metagenome datasets.

From what I understand based on the version uploaded to bioRxiv, there was exploration into its use for microbial genomes. However, it seems this part was excluded from the published version during the revision process (I saw it in peer review comments).

1. Are there any current recommendations for applying COBRA to whole metagenome datasets?
   Is performing contig extension with COBRA before binning currently the best practice, as mentioned in the revision comments?

2. Is it possible to apply COBRA like bewlow workflow (mutiple binning tools followed by binning refinement)

1. contig extension with COBRA
2. binning using multiple binner (such as MetaBAT, MaxBin, CONCOCT, Binsanity etc.)
3. binning refinement (e.g., DAS tool) process.

I would assume performing this process individually for each sample wouldn't raise any problems, but I wonder if you have any experiences or concerns regarding this workflow.

Thank you for taking the time to read my questions.
If there are any misunderstandings on my part, I would appreciate your kind correction.

Regards,

Hello Cobra authors!

Thanks for building this tool.

I keep getting this error of a contig file not being created.

Traceback (most recent call last):
  File "/home/mall0133/miniconda3/envs/cobra/bin/cobra-meta", line 10, in <module>
    sys.exit(main())
  File "/home/mall0133/miniconda3/envs/cobra/lib/python3.8/site-packages/cobra.py", line 1747, in main
    '\t'.join([contig, str(header2len[contig]), summarize(contig), query2current[contig]]) + '\n')
  File "/home/mall0133/miniconda3/envs/cobra/lib/python3.8/site-packages/cobra.py", line 575, in summarize
    b = count_seq('COBRA_retrieved_for_joining/{0}_retrieved.fa'.format(item))  # number of retrieved contigs
  File "/home/mall0133/miniconda3/envs/cobra/lib/python3.8/site-packages/cobra.py", line 482, in count_seq
    a = open(fasta_file, 'r')
FileNotFoundError: [Errno 2] No such file or directory: 'COBRA_retrieved_for_joining/NODE_1990_length_17225_cov_13.396421_retrieved.fa'

Every time I try re-running, it fails at this stage with a FileNotFoundError for a different contig.

This is my Cobra command. I'm using the latest version (version 1.2.3).

cobra-meta -f contigs.fasta -q query_contigs.fasta -c coverage.tsv -m sorted_reads.bam -a metaspades -mink 21 -maxk 127

My query_contigs.fasta file contains about 2300 contig sequences. I've also attached the log file of the run.
log.txt

Any advice on how to fix this error and get Cobra running will be appreciated.

Thanks!

Hi.
This is great tool.
I would like to use cobra for reconstructing complete virome genome-set from my gut metagenomic reads.
Is it possible to provide some small test data (OR mock) on the repo. ?
I am hoping to confirm that it works successfully in my enviroment (ubuntu20, python3.10).

Best,

Thanks for this great tool! However, I have a few questions on input for cobra

1. Now I'm using megahit to assembly (I just know from #3 that it may generate many chimeric contigs, but spades may run out of memory when assembly environmental samples). In megahit intermediate results, it will provide a file named ./intermediate_contigs/k141.contigs.fa, in which there are many small contigs < 200 bp (which will be filtered in ./final.contigs.fa). My question is, which contig file is more recommended to be used as --fasta FASTA input?
2. Is it possible to just use the final.contigs.fa as --query QUERY file, or a filtered version with all contigs longer than given length (i.e. 1000 or 2500 bp)? In another words, can cobra be used before virus contigs annotation and MAG binning?

Regards,
hwrn

Hi Dr. Chen

Thank you for introducing this awesome tool!
What should I do if I have multiple bam files for the assembly built from multiple samples?
Thanks!