PharmUse is a database compiling information on prescription drug consumption, excretion, physicochemical properties, fate and transport parameters, and literature-reported toxicity values for 313 pharmaceuticals. PharmUse integrates data from the nationally representative 2020 U.S. Agency for Healthcare Research & Quality Medical Expenditure Panel Survey (MEPS) Prescribed Medicines File, accessible from here, with the U.S. Environmental Protection Agency's (EPA) CompTox Chemicals Dashboard, drug labels from the U.S. Food & Drug Administration's Drugs@FDA database, the EPA's Chemical Transformation Simulator, and eawag and the University of Auckland's biotransformation database enviPath. The MEPS data are downloaded directly from the provided website and converted to a text file. The physicochemical properties and toxicity values are downloaded directly from CompTox by conducting a batch search for Chemical Name, SMILES, InChi String, InChiKey, Molecular Formula, Average Mass, ToxValDB Details, and Physicochemical Property Values. Individual drug labels for each available administration route were downloaded from Drugs@FDA and manually curated for each pharmaceutical to determine its excretion route. The Chemical Transformation Simulator was used to predict if a given pharmaceutical would hydrolyze and the batch search feature was used to process pharmaceuticals by their SMILES identifier 9 at a time. enviPath was queried for each pharmaceutical to determine its likelihood to biotransform. The script pharmuse_constructor.R was used to load, pre-process, and integrate these data.
PharmFlush is a binomial distribution model that uses data from PharmUse to predict the concentration and mass load profiles of 313 pharmaceuticals in wastewater influent in a sewershed of a given size. PharmFlush was validated by conducting a literature search for pharmaceutical concentration measurements in wastewater influent in the U.S. between 2019 and 2024 and converting them to mass loads. These reported mass loads were then compared against predicted mass loads for sewersheds of the same size as the given study. PharmFlush is run using pharmflush.R.
The script ensemble_exploration.R uses csv files generated in pharmuse_constructor.R and pharmflush.R to compare literature-reported, predicted, and No Observed Effect Mass Loads.
The script get_png_from_pdf.py converts each page of the PDF file generated in ensemble_exploration.R into separate PNG files.
All code is compatible with R 4.4.3 for Windows 11 and RStudio 2025.09.2.418, except the script get_png_from_pdf.py, which is compatible with Python 3.13.9 for Windows 11 and Visual Studio Code 1.107.1.
All packages are R packages that can be downloaded from CRAN except pdf2image (see below).
The script ensemble_exploration.R requires tidyverse Version 2.0.0, ggforce Version 0.4.2, and pdftools Version 3.5.0. Additionally, all_ensembles_100.csv, all_ensembles_1000.csv, all_ensembles_lit_size.csv, all_ensembles_1mil.csv, wrrf_profiles.xlsx, noec_vert.csv, noec_invert.csv, noec_human.csv, lc50_vert.csv, lc50_invert.csv, and lc50_human.csv must be in the working directory to run this script.
The script fix_tox_order.R requires tidyverse Version 2.0.0. Additionally, comptox_search_orig.xlsx, previously_missed.xlsx, and ertugliflozin_dexlansoprazole.xlsx must be in the working directory to run this script.
The script get_png_from_pdf.py requires the package pdf2image Version 1.17.0, which can be downloaded using pip (pip Version 25.0.1 was used in this case). This version of pdf2image requires Poppler Release 24.08.0.0, which can be downloaded for Windows from here as Release-24.08.0-0.zip (for macOS and Linux, see below). On Windows, Poppler must be on the correct path for pdf2image to work. This can be achieved by these steps:
- Extract all files in Poppler zip file and copy the path to the bin folder.
- Press Win + S and type Environment Variables.
- Click on "Edit the system environment variables" and then click the "Environment Variables" button in the dialog.
- Under System variables, find and select the Path variable.
- Click Edit.
- Click New and paste the path to the bin folder (e.g., C:\path\to\poppler\bin).
- Click OK to close all dialogs.
- Open a new Command Prompt (Win + R, type cmd, hit Enter). Type: where pdfinfo If it returns the path to pdfinfo.exe, the setup is successful.
For macOS, Poppler can be installed using Homebrew. To install Homebrew, use the following steps:
- Run the following in your terminal: /bin/bash -c "$(curl -fsSL https://raw.githubusercontent.com/Homebrew/install/HEAD/install.sh)"
- Follow the on-screen instructions.
- Verify Homebrew has successfully been installed by running brew --version in the terminal.
Then Poppler can be installed by running the following in the terminal: brew install poppler
For Linux, Poppler can be installed by running the following in the terminal: sudo apt-get install poppler-utils
In addition to these dependencies, Lit_pred_tox_mass_load_graphs.pdf must be in the same directory as get_png_from_pdf.py for the script to run. This PDF can be generated using ensemble_exploration.R.
The script lit_comparison_all.R requires stats Version 4.4.3, tidyverse Version 2.0.0 (which includes readxl), svglite Version 2.1.3, patchwork Version 1.3.0, cowplot Version 1.1.3, and scales Version 1.4.0. Additionally, the function pharmflush_function.R (which houses the function format of the pharmflush.R model script) and wrrf_profiles.xlsx must be in the working directory to run this script.
The script pharmflush.R requires stats Version 4.4.3, tidyverse Version 2.0.0, svglite Version 2.1.3, patchwork Version 1.3.0, cowplot Version 1.1.3, and scales Version 1.4.0. Additionally, pharmuse.csv must be in the working directory to run this script. pharmflush.R was converted into a function pharmflush_function.R, which requires pharmuse.csv to be in the working directory to run.
The script pharmuse_constructor.R requires tidyverse Version 2.0.0. Additionally, the functions library pharmuse_constructor_functions.R, the MEPS data file MEPS_data_2020.txt, drugs_to_remove.xlsx, excretion_data.xlsx, comptox_search.xlsx, biotransformation_data.xlsx, and hydrolysis_full_data.xlsx must be in the working directory to run this script.
SeweRx has been converted into an R Shiny web app, located here. The zip file shiny_app contains the script app.R and the pharmflush_function.R script (the same as the pharmflush_function.R script provided in the main directory but with a different output format to be compatible with the Shiny user interface). app.R generates the Shiny app and requires pharmflush_function.R and pharmuse.csv to be in the same working directory to work. The same package versions are required to run PharmFlush as outlined above, and the following packages are also required to run the app.R script: shiny Version 1.10.0, bslib Version 0.9.0, DT Version 0.33, tidyverse Version 2.0.0, RColorBrewer Version 1.1.3, and shinycssloaders Version 1.1.0.