ucbd2k/hms-cellcount repository overview

⁠Docker container for generating Growth-rate inhibition (GR) values and fitted curve metrics in RStudio

This docker container provides a platform for processing and analyzing cancer cell line drug-response assays generated by the HMS LINCS Data and Signatures Generation Center. The platform is provided in the form of an RStudio implementation. We also provide an R script that demonstrates the use of the plaform in processing one of the cell count datasets released through the Lincs Data Portal. The processing starts with Level 2 data and ultimately produces Level 4 data.

The script averages cell count data over technical replicates (estimated by intra-cellular ATP levels measured by a CellTiter-Glo® assay) and fits the data to 3-parameter logistic dose-response curves using the growth-rate inhibition (GR) method of Hafner et al. 2016⁠.

See here⁠ for more details on the GR methodology and GR metrics.

The parameters that define the dose-response curves are given as well as other calculated dose-response metrics (e.g. GR50, GR_AOC). The docker uses the "Breast Cancer Density" data generated at the Harvard Medical School HMS LINCS Center as an example.

These data start with cell counts from individual wells (Level 2)⁠. The script Level2_to_Level3_and4.R averages these counts over technical replicates and gives a data frame equivalent to the Level 3 data⁠. It then calculates GR values (analogous to relative cell counts) for the averaged cell counts and fits the data to dose-response curves, giving a data frame of GR values and a data frame of fitted curve metrics equivalent to the Level 4 data⁠. Traditional metrics based on relative cell counts, such as IC50 and Emax, are calculated as well and reported for comparison.

⁠Installation of Docker

Ubuntu: follow the instructions⁠ to get Docker CE for Ubuntu.

Mac: follow the instructions⁠ to install the Stable verion of Docker CE⁠ on Mac.

Windows: follow the instructions⁠ to install Docker Tookbox⁠ on Windows.

⁠Download and run the docker container

Note: You may need to add sudo before the following commands on Linux or Mac

To obtain the docker image and run the container,

docker pull ucbd2k/hms-cellcount:stable

To run the container, execute the following command:

docker run -d -p 8787:8787 ucbd2k/hms-cellcount:stable

If port 8787 is not available, you may need to try another port, for example docker run -d -p 7777:8787 ucbd2k/hms-cellcount:stable.

To start an RStudio session, open a web browser and type http://localhost:8787 in the address bar. Enter rstudio for both username and password.

Note: If localhost does not work (for example if you are using Docker Toolbox for older Mac or Windows systems), use the docker host ip address instead of localhost. Determine the docker host ip with the docker-machine ip default command and substitute this ip address for localhost.

⁠Execute the processing pipeline

After entering the RStudio environment, type the following command in the console:

source("Level2_to_Level3_and4.R")

You can also open the R code by clicking the file named Level2_to_Level3_and4.R from the files panel in the bottom-right of your window. Once you see the code appears in the top-left window, you can click Source at the top of the window.

This script will average the cell count data over technical replicates, converting it from Level 2⁠ data to Level 3⁠ data. It will then fit (GR) dose-response curves to the averaged cell counts and calculate curve metrics, giving Level 4⁠ data. When it is done, the script will show the Level 3 data (averaged cell counts) and Level 4 data (GR values and fitted curve metrics).

⁠Note

The output from this script differs slightly from the released GR metrics data due to some minor modifications in the processing pipeline since the data release.