MUST Screen

Inspiration

Depression has been a severe health problem globally. Antidepressants are among the most commonly prescribed medications in the US, and increasing research demonstrated that genetic variations may influence the effectiveness and side effects for different patients. We have seen our friends suffering from severe depression and it took a long time for them to find out which medication works best for them. The process of trying new drug combinations and suffering from various side effects even worsened their situations.

What it does

MUST Screen stands for Machine-utilized Screening Tool, and it is a free-access screening website for searching for the most suitable antidepressant medication for individuals. A user first selects what condition medication they are looking for and uploads a person’s genome sequence into the website, and the file is fed to our backend MUST Screen model that will run through the entire sequence and compare it with those of other patients’ experiencing both optimal effects and side effects while taking medications that used to treat for the selected condition. A list of medications with side effects will be returned. A user could also search for specific antidepressant medication on our website to receive information on drug-drug interaction regarding that specific medication. The goal of our product is to improve decision-making on choosing antidepressant medications for individuals and clinical professionals, and also personalize medicine.

How we built it

We utilized published Genome-wide association study (GWAS) data related to antidepressant response from Sollis E, Mosaku A, et al, The NHGRI-EBI GWAS Catalog: knowledgebase and deposition resource to compare with an input genome information. We extracted post-treatment expression profiles for drug-drug interaction (DDI) and drug side effect predictions for data analysis in Python. A website was then established using Streamlit.

Challenges we ran into

-Identifying the clinical problem -Finding a suitable dataset -Building the website in 24 hours with no previous experience

Accomplishments that we're proud of

-Building the website with interactive functions with zero experience -Coming up with a meaningful clinical problem and trying to solve it through multi-disciplinary work -Teamwork and communication

What's next for MUST Screen

If possible, a similar workflow can be applied to other diseases for personalized medication. Moreover, some data analysis can be done for side-effect prediction for drugs. We also hope that our website can serve as a database that is beneficial for patients, doctors, and pharmaceutical companies as we can get more information from real patient samples.