UCSC browser track

[maximilianh]

Hi, I work for the UCSC browser and came across your website through searches around webstr. So I added webstr as a track and also your BED file as a separate track to hg38. Is that OK, do you like the display, track names, and the docs page? Here is my own little demo sandbox website to demo it: https://hgwdev-max.gi.ucsc.edu/cgi-bin/hgTrackUi?db=hg38&c=chr4&g=strVar

Should I add it to other genomes? Do you like the container track name "STR Variation", or should it be just "Tandem Repeats" or "Tandem Repeat Variation" ?
Thanks! Max

[hdashnow]

Hi @maximilianh, this is awesome!!! Thanks for doing this.

I have a few thoughts/suggestions:

• I would suggest Tandem Repeats or Tandem Repeat Variation, as the data contains both STRs and VNTRs
• Include the STRchive version number in addition to the date, as we update it ~monthly, so users can easily track changes
• Add a field for the locus-disease association color. For example, some loci in here are included for historical reasons but have been refuted (especially DMD) or have contradictory evidence in the literature. I can add this field to STRchive-disease-loci.hg38.general.bed if that's the file you prefer to pull from. You could then color by it, as you do for ClinGen Validity.
• If you're looking for other population data resources, @bw2 (who did the gnomAD TRs) and co have a nice resource that we use frequently https://trexplorer.broadinstitute.org/
• And finally, would it be useful for us to auto-generate a UCSC track-friendly bed file automatically for each release? We're already generating several other formats, so this likely wouldn't be too difficult. This would allow people to grab the latest.

Let me know what you think!

Warm regards,
Harriet

[maximilianh]

Thanks for the quick feedback!

- I would suggest Tandem Repeats or Tandem Repeat Variation, as the data contains both STRs and VNTRs Done. https://hgwdev-max.gi.ucsc.edu/cgi-bin/hgTrackUi?c=chr7&g=strVar - Include the STRchive version number in addition to the date, as we update it ~monthly, so users can easily track changes Done. See

https://hgwdev-max.gi.ucsc.edu/cgi-bin/hgTrackUi?c=chr7&g=strchive

- Add a field for the locus-disease association color. For example, some loci in here are included for historical reasons but have been refuted (especially DMD) or have contradictory evidence in the literature. I can add this field to STRchive-disease-loci.hg38.general.bed if that's the file you prefer to pull from. You could then color by it, as you do for ClinGen Validity <https://genome.ucsc.edu/cgi-bin/hgTables?db=hg38&hgta_group=phenDis&hgta_track=clinGenComp&hgta_table=clinGenHaplo&hgta_doSchema=describe+table+schema> . Sorry, from these sentences, I don't know how to color what, which color

to assign based on which field in the .bed file. I can use any of your files. I see that there are many files here https://github.com/dashnowlab/STRchive/tree/main/data/catalogs but none of them seems to contain a "locus-disease association" ?

- If you're looking for other population data resources, @bw2 <https://github.com/bw2> (who did the gnomAD TRs) and co have a nice resource that we use frequently https://trexplorer.broadinstitute.org/ OK, I'll try to also make a track from that tsv file. - And finally, would it be useful for us to auto-generate a UCSC track-friendly bed file automatically for each release? We're already generating several other formats, so this likely wouldn't be too difficult. This would allow people to grab the latest. Yes. You can also create a bigBed file, these can be loaded into all

genome browsers, including IGV. You can use the BED9 format, so chrom, start, end, name, score, strand, thickStart, thickEnd, color. You can set thickStart, thickEnd to the same as start and end if you like. Tack on any other fields at the end. Here are the scripts I used to convert your file you can use these, or just change your BED file or just tell me how to color the fields: https://hgwdev.gi.ucsc.edu/~max/harriet/

Let me know what you think!

Looks good!

[maximilianh]

Here is quick trexplorer track: https://hgwdev-max.gi.ucsc.edu/cgi-bin/hgTracks?hgS_doOtherUser=submit&hgS_otherUserName=Max&hgS_otherUserSessionName=trs It's not pretty but should do the job. One day, I should show the histograms right on the click pages... but that's more work...

[bw2]

Nice! Thanks for adding the TRExplorer v2 track.

For easier comparison, what do you think about using the same label style in the WebSTR track as you currently use in the TRExplorer track? (eg. motif x reference repeat count rather than id)?

Also, the TRExplorer item details text includes some inaccuracies, starting from the end:

The References section should say:

Defining a tandem repeat catalog and variation clusters for genome-wide analyses
Ben Weisburd, Egor Dolzhenko, Mark F. Bennett, Matt C. Danzi, Isaac R. L. Xu, Hope Tanudisastro, Bida Gu, Adam English, Laurel Hiatt, Tom Mokveld, Guilherme De Sena Brandine, Readman Chiu, Nehir Edibe Kurtas, Helyaneh Ziaei Jam, Harrison Brand, Indhu Shree Rajan Babu, Melanie Bahlo, Mark JP Chaisson, Stephan Züchner, Melissa Gymrek, Harriet Dashnow, Michael A. Eberle, Heidi L. Rehm
bioRxiv 2024.10.04.615514; doi: https://doi.org/10.1101/2024.10.04.615514

And similarly in the Credits section:
Replace Ben Weiss, Egor Dolzhenko, Michael Talkowski, and the TRExplorer team with Ben Weisburd, Egor Dolzhenko, and the TRExplorer team

In the Population Frequency Data section:

TenK10K description should say 1,925 short-read genomes of European ancestry genotyped using ExpansionHunter
HPRC256 description should say 256 diverse HiFi PacBio genomes from the Human Pangenome Reference Consortium genotyped using TRGT-LPS.

Replace 1,027 samples from the All of Us Research Program with 1,027 HiFi PacBio samples from the All of Us Research Program genotyped using TRGT-LPS

[maximilianh]

Thanks, made all these changes.
https://hgwdev-max.gi.ucsc.edu/cgi-bin/hgTrackUi?g=trexplorer

[bw2]

Thanks @maximilianh ! I like how easy it is to visually compare the two catalog tracks now just by seeing the colors and labels.

Two remaining corrections on the description page:

For TenK10K description, please delete the phrase using STR analysis tools

Also, the Credits section should just say:
"Ben Weisburd, Egor Dolzhenko, and other TRExplorer co-authors."

[maximilianh]

oops, sorry for the typos, fixed now.

[maximilianh]

Hi @bw2 and @hdashnow, I changed the coloring now to color by expected heterozygosity , suggested by Melissa Gymrek. So the features are light grey = no variation, dark grey = no data, and the rest is colored from blue to red by this value. Sounds good?