"Number of clusters formed: 0"

[Former user]

Heia, I am trying to run metabat on my metagenomes. The first one I ran worked ... however for my following ones I get "Number of clusters formed: 0".

Could you help me figure out what I am doing wrong?

I used bbmap/bbwrap.sh to make my bam file. (nohup /homes/nesbocam/bbmap/bbwrap.sh ref=HI.1247.001_spades_scaffolds.fasta in=/groups/edwards/camilla/elisse/HI.1247.001_R1_trim_paired.fastq in2=/groups/edwards/camilla/elisse/HI.1247.001_R2_trim_paired.fastq out=HI.1247aln.bam &)

runMetaBat.sh HI.1247.001_spades_scaffolds.fasta HI.1247aln.bam

The content of my output:

Executing: 'jgi_summarize_bam_contig_depths --outputDepth HI.1247.001_spades_scaffolds.fasta.depth.txt --pairedContigs HI.1247.001_spades_scaffolds.fasta.paired.txt --minContigLength 1000 --minContigDepth 2 HI.1247aln.bam' at Sat May 20 11:16:26 MDT 2017 Output depth matrix to HI.1247.001_spades_scaffolds.fasta.depth.txt Output pairedContigs lower triangle to HI.1247.001_spades_scaffolds.fasta.paired.txt minContigLength: 1000 minContigDepth: 2 Output matrix to HI.1247.001_spades_scaffolds.fasta.depth.txt Opening bam: HI.1247aln.bam Consolidating headers Allocating pairedContigs matrix: 10 MB over 1 threads Processing bam files Thread 0 processing: HI.1247aln.bam Thread 0 finished: HI.1247aln.bam with 29762848 reads and 24098617 readsWellMapped Creating depth matrix file: HI.1247.001_spades_scaffolds.fasta.depth.txt Closing most bam files Creating pairedContigs matrix file: HI.1247.001_spades_scaffolds.fasta.paired.txt Closing last bam file Finished Finished jgi_summarize_bam_contig_depths at Sun May 21 13:50:37 MDT 2017 Creating depth file for metabat at Sun May 21 13:50:37 MDT 2017 Executing: 'metabat --saveTNF saved-HI.1247.001_spades_scaffolds.fasta.depth.txt.TNF --saveDistance saved-HI.1247.001_spades_scaffolds.fasta.depth.txt.distance --inFile HI.1247.001_spades_scaffolds.fasta --outFile HI.1247.001_spades_scaffolds.fasta.metabat-bins- --abdFile HI.1247.001_spades_scaffolds.fasta.depth.txt' at Sun May 21 13:50:37 MDT 2017 [Info] Correlation binning won't be applied since the number of samples (1) < minSamples (10)

Number of clusters formed: 0 Finished metabat at Sun May 21 13:50:40 MDT 2017

Thanks, Camilla

[Don Kang]

You need to supply assembly fasta file as well like this:

runMetaBat.sh <options> assembly.fasta sample1.bam [sample2.bam ...]

[Camilla Nesbø]

Thanks for getting back to me. I am supplying the assembly file: HI.1247.001_spades_scaffolds.fasta - or am I misunderstanding you? (my command: runMetaBat.sh HI.1247.001_spades_scaffolds.fasta HI.1247aln.bam)

[Camilla Nesbø]

My bam file was not sorted - I made a sorted one and it worked.

[Shuangfei Zhang]

@Camilla L. Nesbø. I have the same question like you. you said that you sorted the bam file. can you tell me your specific code about that. my code is following: /home-fn/users/nscc1082/software/samtools-1.3.1/samtools sort -l 0 -o file.bam -O bam -n -T file -@ file.sam. however, my file.bam file is empty.

[Camilla Nesbø]

Hi, I use BBmap I first make the reference /bbmap/bbmap.sh ref=../final.contigs.fa then in the same folder I make the sorted bam file: /bbmap/bbmap.sh in=HI.1247.002_R1_trim_paired.fastq in2=HI.1247.002_R2_trim_paired.fastq out=HI.1247.002mapped.bam bs=bs.sh; sh bs.sh &

[Rob Egan]

Hi Shaungfei,

Your samtools sort command is incorrect in several ways:

1) you must not sort by name! Sort by reference position (i.e. no -n option) 2) -@ takes a number of threads, not a file name 3) -O probably requires 'BAM' (not 'bam'). You don't need to specify it as BAM is the default 4) I believe that samtools sort requires the file to be in a bam format already, not sam

try this: samtools view -Sbu file.sam | samtools sort -@ 8 -o file.bam -

[Shuangfei Zhang]

@Camilla L. Nesbø @Rob Egan Thank you very much! I am a fresh fish. Your advice are good and available. I also have a question to ask you. In my metageomic data, I find a special archea maker gene and the archea may be a new classification. So I want to ask you how to bin a single genome from metagenome precisely. Thank you again. Looking forward to your reply.

[Rob Egan]

Hi Shuangfei, MetaBAT can't just bin 1 genome, it takes the dataset as a whole and bins them all as best it can with all the information available to it. I recommend running MetaBAT on the whole assembly and then find the bin or bins that contain your maker genes for follow-up.

[Shuangfei Zhang]

@Rob Egan Thank you for your reply. Metabat is easy and available. However, I find that the scale of some bins are too big, like 20-40Mb and the maximum of bins are out of control. How do you look at this question? I have no idea. After using some detective softwares, such as CheckM, generally these big bins should be killed.

[Rob Egan]

Hi Shuagfei,

There are a few options that you can try on metabat to vary the sensitivity vs the specificity, but I don't have any advise on where to set the variables off of the default -- every data set behaves differently.

--maxP arg (=95) Percentage of 'good' contigs considered for binning decided by connection among contigs. The greater, the more sensitive.

--minS arg (=60) Minimum score of a edge for binning (should be between 1 and 99). The greater, the more specific.

--maxEdges arg (=200) Maximum number of edges per node. The greater, the more sensitive.

Additionally, I would advise you to try to annotate the large bins. You may find a partial eukaryote or a few very closely related strains in those large bins.

[Yue Lou]

Hi,

I am trying to run metabat but i get the error “[Info] Correlation binning won't be applied since the number of samples (7) < minSamples (10)”.

This is one of my commands:

metabat -a L1_007.jgi.depth.txt -i L1_007_000G1_euk_contigs.fa.gz L1_007_000M1_euk_contigs.fa.gz L1_007_029G1_euk_contigs.fa.gz L1_007_061G1_euk_contigs.fa.gz L1_007_091G1_euk_contigs.fa.gz L1_007_122G1_euk_contigs.fa.gz L1_007_242G1_euk_contigs.fa.gz L1_007_365G1_euk_contigs.fa.gz -o bin -t 6

I have used jgi_summarize_bam_contig_depths to calculate depths for all my samples. I tried both zipped and unzipped fasta files and they gave me the same error. I have also tried to adjust “--minSamples” to 3. When I tried that, it just returns “number of clusters formed: 0”. Could you please help me figure out the problems here?

Thanks!

Clare

[Rob Egan]

It looks like you've done several different assemblies. MetaBAT does not work that way, as it expects to cluster the sequences from a single assembly. It works best when you have multiple samples which contributed to produce that assembly, and each of the BAM files map the reads from a single sample to that common assembly.