Analyses that accompany a manuscript describing the crossmap software.
This repository describes analyses of large datasets that are not included on github. A snapshot of data files - both raw and processed - is available on zenodo (dataset 4287332).
(Repository on github may differ from the zenodo snapshot by a few commits.)
The repository can be used in two distinct ways. It is possible to use the code to start with raw data files and perform a complete analysis, including data-intensive calculations, and produce summary visualizations. Alternatively, it is possible to start with a combination of raw and processed data files, and only generate visualization based on the preprocessed data.
The repository holds code for creating knowledge-bases with crossmap, using these knowledge-bases to process datasets, and visualize results. In order to execute these steps, several setup steps are required. Apart from the setup, the total running-time for these calculations can be several hours.
The crossmap software should be installed in a location that is separate from this repository to avoid any potential file conflicts.
An executable bash script called crossmap should be created at the repository
root to allow the repository to utilize the software. An example script:
#!/bin/bash -l
python3 /path/to/crossmap/crossmap.py $@
The crossmap software will require access to a mongodb instance. See the crossmap documentation.
A list of required packages appears in file vignettes/config.R. Most packages
are available from CRAN while others can be installed through github.
A set of data files must be provided into a data directory. The required
data files include ontology files, gene annotations from several sources, and
other data. A snapshot of all the required files can be obtained from the
data folder in the zenodo dataset.
Raw data must be transferred into crossmap instances. This can be performed
by executing scripts in the prep directory. The README.md in that directory
has additional details on each script.
The vignettes folder contains two Rmd files and a number of other helper
files (R scripts, style sheets, etc.)
To compile the vignettes, navigate to the vignettes folder, start a new R
session, and render the vignettes.
render("CrossmapFigures.Rmd")
render("CrossmapSupplementary.Rmd")
When these vignettes are executed for the first time, they perform
compute-intensive calculations using the crossmap instances and custom R
functions. The combined running time can reach several hours. Outputs from
these calculations will appear in results and cache directories. The
rendered vignettes will themselves consist of pdf files in the vignettes
directory.
When the scripts are executed again, even starting from a clean R session, the vignettes will utilize the cached data and regenerate pdf documents in minutes.
It is possible to use the repository together with precomputed data files and produce summary figures bypassing compute-intensive steps. The running time for this approach should be around a minute.
A snapshot of all raw and data files is available in a zenodo repository
(dataset 4287332). Copy the content of
data, results, and cache directories into the root of the repository.
Summary figures are generated via vignettes. Follow the instructions regarding the installation of R packages and vignette rendering (above).