In vjcitn/YESCDS: Cancer data science activities for CZI YES for CURE collaboration

library(tufte)
# invalidate cache when the tufte version changes
knitr::opts_chunk$set(tidy = FALSE, cache.extra = packageVersion('tufte'))
options(htmltools.dir.version = FALSE)
suppressPackageStartupMessages({
library(BiocStyle)
library(YESCDS)
library(ggplot2)
})

Introduction

The YES for CURE program has a long history of engaging undergraduate and high school students in cancer research activities, mostly in wet-lab applications (@Michel2021).

In 2021, with support from Chan-Zuckerberg Essential Open Source Software^[Bioconductor: Sustaining a Worldwide Community of Genome Data Scientists, EOSS D&I], we began work on a series of presentations related to cancer data science. In the summer of 2022, 11 undergraduate students joined a five-session program.

This brochure addresses our approaches to

recruitment
curriculum authoring using Bioconductor^[https://bioconductor.org], R packages^[https://cran.r-project.org/doc/manuals/R-exts.html], and R markdown^[https://rmarkdown.rstudio.com]
Galaxy-based deployment of student notebooks in jupyterlab
forthcoming enhancements accommodating a broader spectrum of authoring approaches
technical internals involving Helm and github actions

Recruitment

We used the poster in Figure 1 to attract the attention of students in the program. The concept map at the top is produced using coggle.it, following suggestions originating in the Carpentries project.

Recruitment poster by Brittany Michel and Vincent Carey.

Curriculum authoring

We used R markdown to author chapters in four groups:

Counting cancers
Mapping cancer rates
Body parts affected by cancer
Clinical trials: Finding new treatments

The chapters are lodged as vignettes in the R package YESCDS, currently managed at github.com/vjcitn.

A Makefile in inst/scripts of the package arranges the translation of R markdown to ipynb for Jupyter notebook-based exploration, using the AnVILPublish package of Martin Morgan (@Morgan2022).

The YESCDS package is also published as a Github pages site, using pkgdown. See Figure 2.

Pkgdown site for YESCDS, see vjcitn.github.io/YESCDS.

Deployment with Galaxy

Figure 3 shows the entry point to the system based on a Galaxy deployment in the NSF ACCESS Jetstream2 academic cloud. Each chapter can spawn an interactive tool in a private VM for each student.

Face page of yescure.bioconductor.org

Forthcoming enhancements

Any github-resident document in a form that Posit's Quarto^[https://quarto.org/] system can process can be contributed as a curriculum component, eliminating the current essential role of R packages. We would expect content testing to be conducted using a github action. See https://github.com/almahmoud/teachMultiplexImaging for an early glimpse.
YESCDS delivers a "code-free" experience, because all operations of significance have been coded into functions in the package, which are simply executed in jupyter notebook cells. A deployment based on the forthcoming YESCODE package will expose more of the programming and will involve the students in code modifications to change data analysis and visualization operations.