bedGraph to bigwig

Question

[Pegasus]

Hi everyone,

I am trying to visualize my bam.file (RNA-seq) using IGV. Following a combination of Biostar and a recommended step posted by #Matt Shirley,

• I converted bam to bedgraph using bedtools genomecov -ibam
• Generated the indexed sized genome using faidx input.fasta -i chromsizes
• However, using the command line bedGraphToBigWig SAMPLE.bedgraph GENOME_FASTA_INDEX SAMPLE.STRANDRaw.bigwig, generated 0 byte-size bw file.

1. My fasta.file is a small genome that contains 52 Contigs/scaffolds belonging to a one chromosome (Bacteria).
2. Using the original fasta file also generated 0 byte-size bw file
3. Is there any tool to merge the 52 contigs generated by Medusa scaffolder, and is such Merging essential for better downstream analysis' results?

Thank you

[ATpoint]

The genome file is a simple two-column file, first one being the chromosome name and second one the size of the chromosome. Example linked below.

https://github.com/igvteam/igv/blob/master/genomes/sizes/hg38.chrom.sizes

[cmdcolin]

random note: you can easily generate a .chrom.sizes file with "samtools faidx", which generates a .fai file, and I believe you can even use the .fai in place of a .chrom.sizes since the first two columns of .fai are the chrom sizes

[Pegasus]

Actually, I started with the fai.file which generated the same 0byte bw file, and this is why I moved to the genomesize as recommended in several threads.

[Pegasus]

@ATpoint Yes, the resulting genomesize file showed a correct pattern (list of all contigs column1 and their sizes column2). However, although bam > bg was generated successfully, bg > bw resulted 0 byte file

[ATpoint]

My bad, I missed that you used the index file already. Please see my other suggestion.

[ATpoint]

Can you give a head of the bedGraph? I think bedtools needs the -bg flag to output bedGraph rather than its histogram-like default.

[Pegasus]

Here is the head of the bg.file (4 columns)

JAFJXZ01    5   2   934 0.00214133

JAFJXZ01    6   1   934 0.00107066

JAFJXZ01    9   2   934 0.00214133

JAFJXZ01    12  1   934 0.00107066

JAFJXZ01    13  2   934 0.00214133

JAFJXZ01    14  4   934 0.00428266

JAFJXZ01    16  2   934 0.00214133

JAFJXZ01    17  2   934 0.00214133

JAFJXZ01    18  3   934 0.00321199

JAFJXZ01    19  6   934 0.00642398

[ATpoint]

It is five columns, no? And indeed not bedGraph format. Try adding the -bg or -bga option to genomecov.

[Pegasus]

I included the argument -bg in my command line;

bedtools genomecov -ibam < input file name> -g <genome.gff> -bg <output name>

Although I got this error below, the command line resulted a bg file (yes, with 5 columns)

[mii] -g not found! Similar commands: "cc", "sq", "vs"

When I repeated same command with the genome.sizes file instead of gff, I got this error

Genome (-g) files are ignored when BAM input is provided.

Working on Linux

Answer 1

[Shiyi]

Not sure if you problem is solved I think if you could try your approach on a hg38 or mm10 file and see if that work will be a good debug point It can validate if your problem is coming from your code or from the genome you are working on, or both (worst case)

Answer 2

[colindaven]

Save yourself the trouble, and go directly from bam to bigwig with deeptools bamcoverage

https://deeptools.readthedocs.io/en/develop/content/tools/bamCoverage.html