cidgoh / virus-mvp Goto Github PK

Footer was made pretty quickly prior to first official release.

Make it look more professional, and also include link to source code (github) and contact email.

• Somewhere on the visualization, need to add a button that can be used for downloading surveillance report
• Surveillance report will be generated during the genomics workflow when the user uploads an input file

Good place for table, mutation details currently in popup, and more

Toggling clade defining mutations switch on and off, and then resizing the window, resizes the heatmap cells. The bars should stay the same size at all times, with scrollable overflow.

Good place to start investigating would be the clade defining mutations switch callback. There is something different about how the heatmap is first rendered on application launch, outside the callback, because this bug does not happen without toggling the switch.

Missing sample size attr. May be more problems.

e.g,

• Epsilon
  • B.1.427
  • B.1.429

You can select all Epsilon lineages by clicking the header, or individual ones.

Encode binary VOC and VOI attributes in heatmap

Some end users have expressed interest in being able to filter lineages by mutations.

Enhance input options by adding single genome vcf file for high-throughput analysis.

Good to have loading indicators for operations that may take a while

e.g., regenerating the heatmap cells figs when there are a lot of strains or running the pipeline

Should round off alt freqs and think of a better way to summarize mutations instead of listing as many as we can fit into the hover box

Heatmaps are pretty good at visualizing a large number of rows and columns, but our cells are pretty big and resource intensive due to marker annotations, hover boxes, heterozygosity spacing considerations, etc.

We might need to limit the number of rows a user can see, while still allowing user to select from a pool of rows exceeding that number from the select lineages modal. Another consideration is whether we want to make the heatmap cells and y-axis scrollable, and then limit the number of rows visible based solely on performance considerations--not screen real estate. Scrolling is easier than having to select lineages.

Add gvf & surveillance files from VirusSeq data portal for the main release

Encode sample sizes. One idea is to have a column to the right of the heatmap.

Related #56

We will need to display these without colour since frequency is not a factor.

Things to consider for distinguishing single genomes from lineages:

• All black or white filled squares
• Different shaped markers?

Change lineage order of hide lineages, upload file, and rendering error

Refresh--rendered just fine

Started in #44

Things to do:

• Instructions for interacting with submodule
• Instructions for windows users

This is a regression after #118

Open select lineages modal. Do not change anything. Click "OK". Data reloads.

Expected behaviour: data does not reload

Some end users have expressed difficulty parsing the borders on their monitors

After #66 and #75 are settled, we may want to display only a subset of lineages to begin with. With all lineages displayed, the heatmap cells callback takes about 1 second to run, which slows things down a bit.

• We need to have a separate version of COVID-MVP (possibly on website branch) that doesnt have user upload option
• We will still need surveillance report downloading option for the data we will be hosting - VirusSeq Data portal

Bold letters in x-axis == intergenic

Plus == insertion

X == deletion

One idea: thicker borders

Perhaps move it to the top of the page

Investigate the cause of this

N/A could imply we did not look for a function for a mutation or that we are missing data. We should use something more specific, that more strongly implies a function has not yet been discovered for a mutation.

Enhance data input options to accommodate non-GISAID (no metadata) sequences or .fasta sequences and list of ids as .txt file

Linked to #53

Getting slow now with ~700 lineages

Possibilities:

• Greater parallelization of viz generation. Maybe one diagram per heatmap row?
• Improve data parsing efficiency when only slight changes are made, or split up data parsing into parallel code for each lineage

Scrolling to a position along the x-axis may get annoying as the number of mutations increases. Need a way to jump to a position directly.

One possible solution: search fn that allows you to jump to a specific nucleotide position.

Display fns as normal but let users know sites are problematic

Started in #48

When a user deletes a user-uploaded sample, the outer div surrounding the heatmap does not resize as needed.

https://www.nature.com/articles/s41396-023-01368-2

Slack Message

• Cache does not seem to be re-accessed on page reload--fix this
• Some of the other intensive operations could be cached (e.g., heatmap generation)

Currently, data_parser.py is converting VCF to GVF which is then parsed for visualization. Lets move towards integrating nextflow commands now which will require following updates

1. Allow user to upload .vcf or .fasta or .tsv file.
2. Need a check to confirm the file type based on which this nextflow command will be structured

nextflow run nf-ncov-voc/main.nf -profile < conda | singularity | docker > --prefix < unique_user_id > --mode user --input_type < vcf | tsv | fasta > --userfile < user_uploaded_file > --outdir < output directory where files will be generated >

If the space needed by the hoverbox exceeds the number of currently visible rows, the hoverbox will not display. #55 related

1. Hide strain
2. Upload strain with same name
3. Expect: error; Result: success

@anwarMZ What was the message you wanted in the tooltip?

Encode mutations not seen in reference lineages after user uploads sample(s)