What it does
After obtaining the sequencing data of the bacterial strain a patient is afflicted by, the health care practitioner could paste it into the web app, and it will determine if the strain has resistance against Azythromycin, Cefixime, and/or Ciprofloxacin. The practitioner can then prescribe the appropriate medication.
How we built it
We trained a random forest model for each drug, with data obtained from https://www.kaggle.com/datasets/nwheeler443/gono-unitigs/data.
During a previous iteration, we tried to use a single model to classify between all three resistance, hypothesizing that perhaps resistance against one drug can affect its resistance against another. But after training it, we realized that because the number of data points for each antibiotic was vastly different, there was bias towards classifying towards certain drugs. So we decided to train the models separately.
Accomplishments that we're proud of
It's a small and simple project, but it's complete and functional!
What's next for Antibiotic Resistance Classifier
Incorporate other data that was available in the dataset, such as geographic region. Can also extend this to other drugs.
Log in or sign up for Devpost to join the conversation.