dashnowlab / STRchive

Short Tandem Repeat disease loci resource
http://strchive.org/
MIT License
10 stars 2 forks source link

gnomAD plot suggestions #76

Open hdashnow opened 3 days ago

hdashnow commented 3 days ago
  1. Suggest hover over for the x axis: Pathogenic Genotype (%): % of individuals predicted to be affected based on their genotype
  2. gnomAD plot needs a legend for the red diamonds and blue bars. I'm assuming this is the estimate and 90% CI. Alternatively, using more standard error bar plotting style would make a legend unnecessary.
  3. We should either remove plots for the loci we deem to be genotyped unreliable in gnomAD or add some kind of warning so people don’t take them literally. For example, HOXA13 loci aren't actually that common, they're just hard to genotype.
vincerubinetti commented 2 days ago
  1. Because of the charting library we're using, I don't think this is possible without work arounds. Even if it could work, I'd worry that people wouldn't know to hover over the axis label. That's not typical in my experience. Instead, I think I'll put a note about it below the chart.
  2. I'll try to a standard error range. It might be difficult because of library limitations, but I'll at least find some way to make it clear.
  3. Unless I have some clear way to discriminate which plots to keep just from the data, I'll let someone else handle manually removing the plot data for the unreliable ones.