I put my bamcoverage data in IGV,and the range is about 0-100 or 0-50.
But I searched some paper in which the IGV range is very small like 0-1 or 0-5.
So ,Does anyone konw how to make it?
many thanks!!!
Hi @hpc
Do you want the data itself rescaled? If so … that might be possible with whatever tool you used to generate the original file. The tool bamCoverage has this near the top – see the option Bin size in bases.
If you are trying to change the display window scaling instead, that is definitely possible. Check the IGV manual – or you might just be able to find where the zoom in/out options are.
Please clarify more if I am misunderstanding. Screenshots might help.
thanks for your reply!
I want to rescale my data. I bamcoverage my data and bin is 50,other option is defalut. normalizetion is RPKM
Then I put the bw file in IGV, the IGV scale range is about 0-200 ,if I set the range to 0-10 in IGV, the peak is to high.So I guess the problem is in the normalization.
I just tried reset the bamcoverage binsize to 500,and the IGV range is samller (about 0-50),but is there other way？because binsize 500 is too high,is it ok in paper？ Sorry I am fresh in bioimfomatics. 
many many thanks for your help.!
Hi @hpc
Just to be clear:
The “bin” size with bamCoverage is the number of genomic base pairs to create a summary value for. You can think of this as a horizontal-width value – the X axis. The smaller the bin size value, the finer the resolution along the genomic strand will be.
It sounds like you are talking about the vertical-height instead – the Y axis? That is the summary for a bin for the features you are counting. The options to control this metric are under the section Scaling/Normalization method. Maybe try with no normalization, then 1x, then the others?
The help section on the tool form (scroll down) explains with more details, plus this tool is included in many tutorials. Maybe review those results in the tutorial, along with the explanations, to see if that helps?
I probably misunderstood originally about this, I originally thought you were talking about the width of bins (number of genomic bases covered per bin). Instead, you are talking about the height of the bins? The “count” per bin? Did the authors also use the RPMK scaling? The Y axis should be labeled if this is in a publication.
It is hard to make more guesses – you are still welcome to share some data if you have more questions.
Thanks for your relpy!
upper is my IGV track,the range is 0-275
and this two are the IGV from two published paper
I just wonder why their range is small .
I run the same normalized way as this two paper ,RPKM ,bin=10。
many many thanks for your help！So thanks for your patience!!!
@jennaj
Thanks for your relpy!
upper is my IGV track,the range is 0-275
and this two are the IGV from two published paper
I just wonder why their range is small .
I run the same normalized way as this two paper ,RPKM ,bin=10。
many many thanks for your help！So thanks for your patience!!!
Hi @hpc
Let’s go from the bottom up:
The third graphic seems to be compressed on the X-axis (genome), and the bit at the bottom looks like a reference gene. The Y-axis is the same for all – but it is the relative values within each sample that matter. The overall shape between the first two samples is about the same, so maybe they are “similar”. I’m don’t think more can be said without knowing more about the exact data/experiment.
The second graphic has values that vary less within each sample, and different mix-max values even between the two samples included … and is sort of a good demonstration about why comparing RPKM values directly between samples, even within the same study, isn’t so meaningful. RPKM values within a single sample can be compared within that same sample using relative differences only. This author had to scale one sample almost twice as large (Y-axis) to generate their visual comparison. Presumably the X-axis (genome/bins) is the same scale for both.
The first graphic – your data – has both larger value counts, and more difference between the low and higher values across the genome than the other two graphics. “Why” could be for a lot of reasons for both – but I will say that there are statistical processes that can address outlier data. If reads are piling up with high coverage in a particular regions, that can bias the result interpretation in lower coverage regions that may be of more interest. That can be applied during laboratory manipulations or post-processing of reads (or whatever feature you are examining). Check publications for the protocol you are following, it is probably covered.
Maybe that helps?
very thanks for your reply and patience！
I have sovled my question!
I just change the normalization method from RPKM to 1×，and the two tracks in following figure are same sample ,same genome region but different normalization method.First track is RPKM and the range is 0-300.Second track is 1× and the range is 0-25,which is what I want!(actually not I want ,my BOSS want!)
many many many thanks !!! I have learned a lot ! thank you again!
1 Like