Data join between 3D model (coord system) and values

[sehilyi]

(still actively updating notes here...)

Goal

• We want to enable users to provide a 3D model (the 3D coordinate system) in addition to their data.
• We need to find a way that harmonizes with the current Goslign grammar (i.e., it shouldn't change the existing grammar significantly).
• We want to find a way that is general enough so that the other layout option (i.e., space-filling layout) can be applied with the same approach.

Example Datasets (Simplest)

Model

x	y	z	chr	coord
20.713	-14.117	-5.7	chr f	3000000
20.536	-13.494	-4.919	chr f	3100000
20.381	-13.234	-3.995	chr f	3200000

Values

chr	position	value
chr f	3000000	0.1
chr f	3100000	10.4
chr f	3200000	15.2

Grammar

Option 1 (join data transform)

{
  views: [{
    layout: 'spatial',
    tracks: [
      { data: { type: 'csv', url: '...' }, // track 1
        spatial: {
          x: "x",
          y: "y",
          z: "z",
          chr: "chr",
          coord: "coord"
        },
        dataTransform: [{
          type: 'join', url: 'https://raw.githubusercontent.com/dvdkouril/chromospace-sample-data/main/gosling-3d/yeast_model.csv', ...
        }],
      }
      { ... }, // track 2
    ]
  }]
}

Option 2 (view-level 3D model definition)

{
  views: [{
    layout: {
      type: 'spatial',
      model: {
        type: 'csv',
        url: 'https://raw.githubusercontent.com/dvdkouril/chromospace-sample-data/main/gosling-3d/yeast_model.csv',
        x: "x",
        y: "y",
        z: "z",
        chr: "chr",
        coord: "coord"
      },
      tracks: [
        { data: { type: 'csv', url: '...' }, ... }, // track 1
        { data: { type: 'csv', url: '...' }, ... }, // track 2
      ]
    }
}]}

Challenges

• The potential mismatch in terms of bin sizes and positions between the 3D model and values (e.g., 2kbp in the 3D model vs. 3kbp in values).

Example 3D model dataset:

• https://raw.githubusercontent.com/dvdkouril/chromospace-sample-data/main/gosling-3d/yeast_model.csv
  • https://gist.githubusercontent.com/sehilyi/29d1cfed56da3ed37370f31a508c8024/raw/9ab733bff25c4c539d86451dfb9d03d4e241d1ad/yeast_model.values.csv
• https://raw.githubusercontent.com/dvdkouril/chromospace-sample-data/refs/heads/main/gosling-3d/Stevens-2017_GSM2219497_Cell_1_model_1_chr_f.csv