Annotations and data layers

Starting with version 1.4, R2DT can visualise additional annotations as data layers on top of the secondary structure diagrams.

By default, R2DT displays the expected accuracy of the alignment between the query sequence and the template. The accuracy is calculated by the Infernal software as posterior probabilities (see page 110 of the Infernal User Guide for more information).

Alignment accuracy

For example, the following diagram shows the secondary structure of a SAM riboswitch RNA sequence from Bacillus subtilis visualised using the RF00162 Rfam template:

The colored circles highlight the hairpin loop region where Infernal is not 100% confident in the alignment.

Hovering over the nucleotides shows the alignment accuracy (posterior probability) values. Similar to Infernal #=GR PP lines, the posterior probability p is encoded as 11 possible characters from 0 to 10:

• 0.0 ≤ p < 0.05 is coded as 0

• 0.05 ≤ p < 0.15 is coded as 1 (… and so on …)

• 0.85 ≤ p < 0.95 is coded as 9

• 0.95 ≤ p ≤ 1.0 (shown as * in Infernal output) is coded as 10.

The nucleotides are colored based on a color-blind friendly palette from a paper by Bang Wong. The default colors can be found in the colorscheme.json file and are shown below:

• 0-5 or <55% accuracy

• 6 or ≥55-65% accuracy

• 7 or ≥65-75% accuracy

• 8 or ≥75-85% accuracy

• 9 or ≥95% accuracy

• 10 and * are shown with white background.

Example annotation file

The annotations can be provided in a tab-separated format where the first two columns are called residue_index and residue_name. The number of rows in the file (not counting the header line) and residue names must match the query sequence.

The third column (in this case posterior_probability) can have an arbitrary name matching the colorscheme.json.

The following example is shortened for display purposes but a complete TSV file is available for reference:

residue_index	residue_name	posterior_probability
1	U	10
2	U	10
3	C	10
...
50	G	10
51	G	10
52	U	10
53	G	9
54	U	9
55	A	9
56	A	8
57	U	8
58	G	7
59	G	8
60	C	8
61	G	8
62	A	8
63	U	8
64	C	8
65	A	8
66	G	8
67	C	8
68	C	8
69	A	7
70	U	7
71	G	9
72	A	10
73	C	10
...
117	A	10
118	A	10

Example color scheme file

The colors for each possible posterior_probability value from the TSV file are configured using a colorscheme.json file, for example:

{
    "coloring": {
      "posterior_probability": {
        "label": "Alignment accuracy",
        "values": {
          "0": "rgb(213, 94, 0)",
          "1": "rgb(213, 94, 0)",
          "2": "rgb(213, 94, 0)",
          "3": "rgb(213, 94, 0)",
          "4": "rgb(213, 94, 0)",
          "5": "rgb(213, 94, 0)",
          "6": "rgb(230, 159, 0)",
          "7": "rgb(240, 228, 66)",
          "8": "rgb(86, 180, 233)",
          "9": "rgb(0, 158, 115)",
          "10": "rgb(255, 255, 255)",
          "*": "rgb(255, 255, 255)"
        }
      }
    }
}

Note that the header of the third column of the annotation file must match the corresponding field of colorscheme.json.

Example R2DT commands

You will need 3 files:

• query.fasta - a query sequence rna_id in fasta format,

• annotations.tsv - per-residue annotations in TSV format,

• colorscheme.json - RGB colors for each annotation category in JSON format.

1. Visualise secondary structure using R2DT:

   # predict and visualise secondary structure using templates
   r2dt.py draw <query.fasta> <output_folder>
   
   # alternatively, visualise provided secondary structure without a template
   r2dt.py templatefree <query.fasta> <output_folder>
   

2. Generate additional SVG file colored using custom annotations:

   python3 /rna/traveler/utils/enrich_json.py --input-json <output_folder>/json/<rna_id>.colored.json \
   --input-data <annotations.tsv> --output <output_folder>/json/<rna_id>.enriched.json
   
   python3 /rna/traveler/utils/json2svg.py -p <colorscheme.json> \
   -i <output_folder>/json/<rna_id>.enriched.json -o <output_folder>/svg/<rna_id>.enriched.svg
   

   The final result will be found in <output_folder>/svg/<rna_id>.enriched.svg.

These commands add metadata from the TSV file to an RNA 2D JSON Schema file using the enrich_json.py script and then generate an SVG file using json2svg.py (both scripts are part of the Traveler software and are available in the R2DT Docker container).

Displaying other types of data

It is possible to visualise other residue annotations, such as the location of single nucleotide polymorphisms (SNPs), conservation, SHAPE reactivity, location of RNA protein binding sites, and other information.

To visualise custom data:

• the annotations should be provided in the third column of the TSV file,

• the posterior_probability field of the colorscheme.json file should match the name of the third column of the TSV file,

• the values field of colorscheme.json should match the possible annotation values.