Welcome to Birnam Wood, a program aimed to reduce the number of parameters required to make accurate kinetic rate constant predictions.
The following packages must be installed prior to running Birnam Wood
pip install -U pandas
pip install -U numpy
pip install -U scikit-learn
Start by cloning the reposititory by using Console or the Github Application.
Once installed, make sure the dataset files are accessible. First run all cells
within nested_cross_validation.ipynb this is a lengthy process as multiple
models will be developed and evaluated. The outcome will be four spreadsheets,
one for dataOff and dataOn which contain each model containing the MSE value for
each model, as well as the number of parameters. While DataOffJ and dataOnJ
contain the resulting J score values. The first two columns of each dataset
represent an integer key used to identify particular models.
Finally, an output will show the best models based off of the Jscore value. Use the model that you see fit.
The next step is to gather data for your protein. More information regarding this process can be found from our wiki page.
Finally, once a csv is created with the proteins and its respected values, open
up Birnam_Wood.ipynb and replace the name of our spreadsheet, lanM.csv
and run the program to recieve a prediction. A sample has been provided for
reproducibility.