Combinatorial Drug Screen

Code for processing and analysing the combinatorial screen data presented in the manuscript "High-throughput profiling of drug interactions in Gram-positive bacteria” (https://www.biorxiv.org/content/10.1101/2022.12.23.521747v1).

All drug-drug interactions (with their effect size and adjusted p-value) detected with this pipeline are listed in Supplementary Table 2 (antibiotic screen conducted in S. aureus Newman and DSM 20231, B. subtilis 168 and S. pneumoniae D39V) and Supplementary Table 6 (non-antibiotic drug screen in S. aureus DSM 20231). https://www.biorxiv.org/content/10.1101/2022.12.23.521747v1.supplementary-material

The repository tree has the following structure:

• data: unprocessed optical density measurements
• dataprep folder: R scripts for pre-processing of the raw screen data
• qualitycheck directory: scripts for quality control
• analysis: high-level analysis of the screen data, computation of interaction scores, p-values and network analysis
• phylogeny: code and data for the phylogeny analysis presented in Fig. 1e
• growthrates: growth rates from fitting a Baranyi model (Methods)

Note that here we use the following codes for the 4 Gram-positive bacterial strains:

Code	Strain
K	S. aureus Newman
M	S. aureus DSM 20231
D	B. subtilis 168
P	S. pneumoniae D39V

The main branch contains the data analysis pipeline for the drug combinatorial screen in Gram-positive bacterial species. In order to run the pipeline in the SLURM cluster environment:

sbatch pipeline.sh datadir batchnumber strain

For example, provided that the data is in and you wannt to process the data from batch 2 (strain K), submit the job as follows:

sbatch pipeline.sh data/screen_K 2 K

Furthermore, the data can be analyzed at the strain level, so that the significant interactions are identified separately for each strain. First run permutation test in parallel:

sbatch permtest.sh datadir strain

For example if one wants to process all of the data for strain K:

sbatch permtest.sh data/screen_K K

Next run plot.sh script to generate visualizations of individual drug combinations:

sbatch plot.sh datadir strain
# strain K
sbatch plot.sh data/screen_K K