Purpose

Our plot has limited room to display genes in a standard figure, and some users have suggested that it might be nice to intelligently choose which elements are shown, as not all gene entries are of equivalent broad interest (see: gencode list of biotypes). For example, the PheWeb nearest gene annotator uses a specific subset of options.

The new LZ filtering feature can be used to render only user-selected biotypes.

Requirements:

[x] The portaldev API server will need to expose gene biotype field
[x] A display_options dropdown menu will need to be added to the genes track with a sample biotype field. We should identify what the most relevant groups of genes to show are- ideally 1-5 preset groups.
[x] Add new "in" and "match" filter functions to allow search-like behavior with filters.
[x] Modify genes track rendering to perform layout on filtered dataset. (instead of leaving space for the hidden elements as it currently does)

Rather than allowing the user to drag and reorder genes, they would thus be selecting the data to render based on the intended meaning/ most common use case.

abought commented 4 years ago

Goncalo suggests a similar list to pheweb.

protein_coding category as definite must. Excluding pseudogenes would be nice; might want to include mirnas. ig c/d/g good to include. For RNA, keep Mt_tRNA, Mt_rRNA.

A long term stretch goal would be a checkbox list for fine grained control.

For UI, option lists would be called "all" or "genes". (excluding micro, non coding, etc)

IG_*_GENE, protein-coding, TR_*_GENE, rRNA, Mt_?RNA 

pjvandehaar commented 4 years ago

For reference, here are the counts on hg38 for gencode v34:

19959   protein_coding
144 IG_V_gene
18  IG_J_gene
14  IG_C_gene
37  IG_D_gene
106 TR_V_gene
79  TR_J_gene
6   TR_C_gene
4   TR_D_gene
22  Mt_tRNA
2   Mt_rRNA
52  rRNA
8   ribozyme
# Total: ~20,000

16899   lncRNA
2212    misc_RNA
1901    snRNA
1881    miRNA
1061    TEC
943 snoRNA
49  scaRNA
5   sRNA
1   vaultRNA
1   scRNA
# Total: ~25,000

10175   processed_pseudogene
2629    unprocessed_pseudogene
927 transcribed_unprocessed_pseudogene
501 rRNA_pseudogene
498 transcribed_processed_pseudogene
188 IG_V_pseudogene
137 transcribed_unitary_pseudogene
97  unitary_pseudogene
42  polymorphic_pseudogene
33  TR_V_pseudogene
18  pseudogene
9   IG_C_pseudogene
4   TR_J_pseudogene
3   IG_J_pseudogene
2   translated_processed_pseudogene
1   translated_unprocessed_pseudogene
1   IG_pseudogene
# Total: ~15,000

abought commented 4 years ago

Thanks Peter- these counts could be useful if we decide to add more than two groups of filters. (which I think might become a real user request after release)

I'll push a new release to my.locuszoom.org tomorrow AM with this feature, so we can get real user feedback!

statgen / locuszoom

Allow users to filter genes track by biotype #199

Purpose