Contradiction between Total Counts and Mean Counts in tdf coverage

[Cobrinik Lab]

I have loaded a set of .tdf files split into treatment and control groups. I expect higher coverage of some particular coordinates in the treatment samples. I find this to be the case when zoomed out and hovering the pointer to show "Total Counts". However, when I zoom in and the hovertext window changes to display "Mean Counts", not only is the pattern reversed, but the mean count values shown are frequently far higher than the total count values displayed at a lower zoom level. I can't find documentation of these two count values in the igv user guide and I have already investigated the downsampling preference in View > Preferences > Alignment referenced on biostars here. I don't think checking or unchecking downsampling has any effect on these displayed values. Any help would be greatly appreciated.

[James Robinson]

Its a little difficult for me to follow what you are describing.   Are you loading just the .tdf files, or are you also loading the associated bam files and using the .tdf as a pre-computed coverage track?

Some screenshots might help.

What version of IGV are you using?

On Mon, Jun 25, 2018 at 2:02 PM, Cobrinik Lab <cobri...@gmail.com> wrote:

I have loaded a set of .tdf files split into treatment and control groups. I expect higher coverage of some particular coordinates in the treatment samples. I find this to be the case when zoomed out and hovering the pointer to show "Total Counts". However, when I zoom in and the hovertext window changes to display "Mean Counts", not only is the pattern reversed, but the mean count values shown are frequently far higher than the total count values displayed at a lower zoom level. I can't find documentation of these two count values in the igv user guide and I have already investigated the downsampling preference in View > Preferences > Alignment referenced on biostars here. I don't think checking or unchecking downsampling has any effect on these displayed values. Any help would be greatly appreciated.

--

---
You received this message because you are subscribed to the Google Groups "igv-help" group.
To unsubscribe from this group and stop receiving emails from it, send an email to igv-help+unsubscribe@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/igv-help/700ee235-f721-47f5-8999-a8f67cff2edf%40googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

[Cobrinik Lab]

Hi James,

Sorry for the confusing description. We're using IGV version 2.4.10 Hopefully these screenshots help clarify.

--

---
You received this message because you are subscribed to a topic in the Google Groups "igv-help" group.
To unsubscribe from this topic, visit https://groups.google.com/d/topic/igv-help/7ipyZP88Jl0/unsubscribe.
To unsubscribe from this group and all its topics, send an email to igv-help+unsubscribe@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/igv-help/CACOP%2Bps41r95cYCTmkAv4a%3DYpsYwSNT3dyQnyK%3D8LvsSo_pGDA%40mail.gmail.com.

[James Robinson]

It looks like you are also loading "bam" files,  you do not see "total count" from a TDF file alone.    When viewing coverage from this many files I suggest you NOT load the bam file if you've already computed it via a TDF.

In your IGV preferences to you have "normalize coverage data (.tdf files only)" checked? 

To view this discussion on the web visit https://groups.google.com/d/msgid/igv-help/CAGRUf%3D-AtHOSgsUSbH7q9rSkQkLwo9aaMj-W4F-%3DsK-EzdEDMQ%40mail.gmail.com.

[James Robinson]

The setting is on the "Tracks" tab  (view > preferences > tracks).

[Cobrinik Lab]

Thanks for following up,

We did try adjusting the "normalize coverage data (.tdf files only)" setting. I should be clear that we're having no issue with tdf display, only with bam files. We're interested in viewing the specific sequences of individual reads, so we're loading bam files instead of tdf.

sorry for the miscommunication. In tdf we see patterns of read counts which are in agreement with the .bam mean count values shown above. At this point we're just wondering what the total count values displayed when inspecting .bam at higher resolution represent.

We want to be sure that we are justified in describing a difference in read counts between the treatment and control groups and that mean count values (but perhaps not total count values) are a valid measure of this.

Thanks,

To view this discussion on the web visit https://groups.google.com/d/msgid/igv-help/CACOP%2Bpvn5eB%3DHDV4%3DjJRL7ku2kQRE8oE6%2ByrDYTSE0o7m3G-qA%40mail.gmail.com.

[James Robinson]

When viewing coverage from BAM files at base level total count is the total number of alignments that cover that base.   Downsampling does not apply here,  reads are counted prior to downsampling.  

The signal from a TDF file should correlate,  but there can be differences particualrly if you have clicked "normalize coverage".  TDF files use bins (windows),  by default 25 bp in width.   The lowest resolution signal in a TDF file is the average count over the window.   As you zoom out these windows are combined to give a signal at the resolution you are viewing, that is the base pair width corresponding to a pixel.   If "mean" is used as the window function (the default),  this will normally flatten peaks.   

If normalize coverage is checked the signal in a TDF file, as described above,  is multiplied by  (1e6 / total # of alignments),  where total # of alignments is the total # of mapped reads in the bam file.  The purpose of this transform is to compare bam files whose total coverage might differ.   This is a very crude normalization obviously.    The normalize coverage option has no effect on the total count value computed directly from the bam file.

To view this discussion on the web visit https://groups.google.com/d/msgid/igv-help/CAGRUf%3D825pERezVBt1W0uVQ8c-QJQfd8b-MGObZ%2BuxFPfz8iGQ%40mail.gmail.com.