PLCO

PLCO is a large NCI trial (n=155,000) that, since 1993 evaluated screening for early detection of prostate, lung, colon, and ovarian cancer. In 2021-2022, summary GWAS data (no individual data!) was made available in the public domain with an Application Programming Interface (API). This creates the opportinity to facilitate data retrieval from a variety of programming languages and analytical environments by developing the corresponding Software Development Kits (SDK).

This is not the data portal

You could develop a data portal using PLCOjs and other libraries mediating access to one or more data sources of interrest. Instead, PLCOjs is a JavaScript SDK developed to provide programatic interoperability with PLCO's public APIs. This developemnt effort was pursued by modularizing PLCO's Data portal graphic plotting and data retrieval components. The rationale is that PLCO data analytics might be integrated with a variety of GWAS data processing applications. This integration is overviewed in the tutorial^[YouTube], the suggested starting point to uderstand PLCOjs.

FAIR

Let's find out how Findable, Accessible, Interoperable and, most defining, Reusable, are the data objects generated by PLCOjs:

If I place my data in a versioned host as in jonasalmeida.github.io/pub/PLCOpca2.json ...

... would I be able to also recover the interactive plot with a URL shortner, such as bit.ly/plcopca?

Give it a try!

Note the plco.js library is also loaded in this page. You can experiment with different plots from the devTools by copy/pasting the example code snippets to the console. The one below was created with:

plco.plot.qq2('plot', [{phenotype_id:3080, sex:'female', ancestry:'east_asian'}, {phenotype_id:3080, sex:'female', ancestry:'european'}, {phenotype_id: 3550, sex:'all', ancestry:'east_asian'}])

Plotting by population stratification See also Plot Documentation. The primary use of this panel is to generate multiple plots for each statification (Manhattan, PCA and QQ). For exmples see this presentation first. The first textarea contains a collection of phenotype id, sex, and ancestry inputs to be plotted. It is initally empty, and you may add more inputs by filling out the input boxes as needed, and pressing the Add button. Finally, press Submit to see your plots! Do note that manhattan plot only supports 1 or 2 inputs, while qq and pca plot supports more than 2 inputs. [ ] Enter Phenotype ID: ^* Enter Sex: ^* Enter Ancestry: ^* Add Reset Submit

Demostrating support with the Epi-Plotly library. Pros: Great for static plots; faster due to not having to fetch API data; no JavaScript code needed; Cons: As of now, the json file must be hosted; https://episphere.github.io/plco/plot.json
Steps to recreate:

1. Set the to_json parameter to true in any of the graphing methods. E.g.: plco.plot.qq('', 3080, 'female', 'east_asian', true)
2. Call it and then parse the results using JSON.parse(...)
3. Pass the JS object into plco.downloadJSON(your_obj)
4. Host the downloaded *.json file anywhere that supports raw files. E.g.: Github, Google Drive, Your Web Server

Tips: You can add Plotly event listeners by querying the div with js-plotly-plot class nested within the epi-plotly tag.
An alternative to Epi-Plotly library. Insert this into the header:
<script src="https://episphere.github.io/plco/plcoHTMLElements.js"></script>

3 plotting tags available:

• <manhattan-plot></manhattan-plot>
• <qq-plot></qq-plot>
• <pca-plot></pca-plot>

All three plotting tags have the following attributes available: data-inputs, custom-layout, custom-config
Example: (Notice that for data-inputs each object is separated by a semi-colon)
<manhattan-plot data-inputs="[{phenotype_id: 3080, sex: female, ancestry: european}; {phenotype_id:3080, sex: female, ancestry: east_asian}]" custom-layout="plot_bgcolor:#d3d3d3;paper_bgcolor:#fff" custom-config="scrollZoom:false" ></manhattan-plot>
The manhattan-plot have additionally chromosome=number and p-val-nlog-min=number attributes. -->

1. 🐛