Can't see colors for ENCODE bigBed on UCSC browser

[paul-sud]

Hello,

We are using gemBS version 3.5.1 and the IHEC standard configuration. When loading the ENCODE-style CpG bigBed produced during gemBS extract into the UCSC browser as a custom track, the colors for all the CpG sites are the same, even with itemRgb=on and different combinations of settings for maxItems, visibility, and type (e.g. bigBed 9 + 5).

However, after manually running bedToBigBed on the ENCODE-style BED file produced by extract, like this:

bedToBigBed -as=bed9_5.as -type=bed9+5 input.bed chrom_sizes.tsv output.bigbed

then the colors are visible on UCSC. The .as file above is from this repo: https://github.com/heathsc/gemBS/blob/master/gemBS/etc/bed9_5.as

Is there some other UCSC setting that is required to see the colors? Or perhaps some issue with the bigBeds produced during extract that prevent the colors from being rendered? Any guidance would be appreciated.

Thanks,

Paul

[heathsc]

Thanks for the report. The documentation for the bigbed format is very scarce and it is possible that some aspect of the summary data is not being produced in the way the ucsc browser is expecting it. The bulk data should be correct, which you can verify with the bigbedtobed tool. I will take another look at the bigbed production and see if I can find out what is required to trigger the colours.

Simon

On Mon, 15 Mar 2021, 20:50 Paul Sud, @.***> wrote:

Hello,

We are using gemBS version 3.5.1 and the IHEC standard configuration https://github.com/heathsc/gemBS/blob/master/gemBS/etc/gemBS_configs/IHEC_standard.conf. When loading the ENCODE-style CpG bigBed produced during gemBS extract into the UCSC browser as a custom track, the colors for all the CpG sites are the same, even with itemRgb=on and different combinations of settings for maxItems, visibility, and type (e.g. bigBed 9 + 5).

However, after manually running bedToBigBed on the ENCODE-style BED file produced by extract, like this:

bedToBigBed -as=bed9_5.as -type=bed9+5 input.bed chrom_sizes.tsv output.bigbed

then the colors are visible on UCSC. The .as file above is from this repo: https://github.com/heathsc/gemBS/blob/master/gemBS/etc/bed9_5.as

Is there some other UCSC setting that is required to see the colors? Or perhaps some issue with the bigBeds produced during extract that prevent the colors from being rendered? Any guidance would be appreciated.

Thanks,

Paul

— You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub https://github.com/heathsc/gemBS/issues/83, or unsubscribe https://github.com/notifications/unsubscribe-auth/AAY4654EANO64GCXEAG5DZDTDZQG7ANCNFSM4ZHGYL5Q .

[paul-sud]

Thanks for offering to take a look Simon, it would be great if there was a way to make the coloring work.

[paul-sud]

Just an update: we are able to see the colors when using a trackHub. The relevant portion (I think) is this:

track test
compositeTrack on
type bed 3
visibility full

    track test_cpg
    parent test on
    type bigBed 9 +
    visibility dense
    scoreFilter 0:1000
    itemRgb on

        track test_cpg_concrete
        parent test_cpg on
        type bigBed 9 +
        ...

Nothing surprising in there, I'd guess there are probably some discrepancies in how custom tracks are handled vs. track hubs.

[heathsc]

Thanks, so it looks like I might just need to change the header.

Simon

On Thu, 18 Mar 2021, 20:56 Paul Sud, @.***> wrote:

Just an update: we are able to see the colors when using a trackHub. The relevant portion (I think) is this:

track test compositeTrack on type bed 3 visibility full

track test_cpg
parent test on
type bigBed 9 +
visibility dense
scoreFilter 0:1000
itemRgb on

    track test_cpg_concrete
    parent test_cpg on
    type bigBed 9 +
    ...

Nothing surprising in there, I'd guess there are probably some discrepancies in how custom tracks are handled vs. track hubs.

— You are receiving this because you commented. Reply to this email directly, view it on GitHub https://github.com/heathsc/gemBS/issues/83#issuecomment-802245521, or unsubscribe https://github.com/notifications/unsubscribe-auth/AAY4655JZ66H3R3EOJS6HY3TEJLFNANCNFSM4ZHGYL5Q .