Run a webserver that will batch process dose-response data using the US EPA's benchmark dose modeling software (BMDS). Under the hood, this web application uses the BMDS Python interface.
To use the webserver, specify a new BMDS Job using a input file formatted like this.
{
"dataset_type": "D",
"bmds_version": "BMDS2601",
"datasets": [
{
"doses": [0, 1.96, 5.69, 29.75],
"ns": [75, 49, 50, 49],
"incidences": [5, 1, 3, 14]
},
{
"doses": [0, 1.96, 5.69, 29.75],
"ns": [75, 49, 50, 49],
"incidences": [0, 0, 11, 27]
}
]
}
You'll receive a Job ID back, and you can check the website to see when the job is complete. Then, results can be downloaded. Results include:
- The dfile created for each model
- The outfile created for each model
- The parsed output file for each model
- A recommended best-fitting model, using guidance from Wignall et al. 2014 (optional)
To run your own BMDS webserver, you'll need a Windows server. Deployment steps have been documented using a Windows Server 2012 and IIS 8.5; it may be possible to run in other environments.