I have downloaded several ChIP-seq data (histone marks) and I'm trying to plot their coverage over the TSS
Since every file comes from a different sequencing experiment the sequencing depth varies from one sample to another. Therefore, what I'm trying to do is scale all the samples by a specific value.
The deeptools package offers that via the bamCompare --scaleFactor function. Is it sensible to calculate the mean coverage for each sample and then apply a scaling factor to normalise all the bam files to each other?
Moreover, I'm having trouble parsing the matrix file from the computeMatrix function into R. If anyone is familiar with deeptools any explanations with regard to what the matrix files represent would be much appreciated
bamCompare is used when you have two bam files. For ChIP-seq there is normally the ChIP sample and the so called input, bamCompare is used to compute the log2ratio or the difference. If you have for all your histone marks the ChIP and input, you can use bamCompare which will produce normalized bigwig files
If you don't have an input, you can use bamCoverage which can normalize the values using two methods: normalizeTo1x and normalizeUsingRPKM. Either method will produce output that is comparable across samples.
You can write to the deepTools mailing list to get support on the matrix issue (firstname.lastname@example.org)