Project:	Using Biological Passports to Detect Performance-Enhancing Drugs
Disciplines:	Computer Science, Biology, Physics
Mentor:	Charles Steinhardt, Professor of Physics, (PMA), csteinhardt@missouri.edu, Phone: 609-672-9866
Background:	NOTE: This project is being offered by a Caltech postdoc alum and is open only to Caltech students. The project will be conducted at the University of Missouri in Columbia, Missouri. This is one of several projects available at the University of Missouri this summer, and we expect that in total 6-8 students (some from Caltech and some from elsewhere) will come to Columbia, MO. We have been running a similar program for a decade, previously in Copenhagen. Other projects in physics with different mentors will also be available. We hope to finalize who will be coming by mid-January, so that there will be plenty of time to both write a SURF proposal and find housing.
Description:	Nearly every sport bans a set of performance-enhancing drugs, and professional athletes who can take them and avoid being caught are able to gain a significant (and unfair) advantage over clean athletes. It is often difficult to directly find banned substances in test samples, in part because there is something of an arms race with new drugs or masking agents constantly in development. One current anti-doping technique relies on a biological passport, in which various biomarkers are monitored over time. An athlete that shows an abnormal progression, such as doubling their red blood cell count overnight without previously having shown a large population of red blood cell precursors, could be assumed not to have gained those red blood cells naturally, even if it is not known precisely what drug or technique was used. However, a clever cheater will be far subtle, so what is needed is the ability to look at a biological passport and determine whether the athlete’s progression is sufficiently anomalous that it cannot have occurred naturally. Making this problem more difficult is that most world-class athletes are already outliers, even if they are not taking banned substances. We have access to the results of a study in which biological passports were taken both for athletes using performance-enhancing drugs in various ways and using advanced but legal training techniques. The goal of the project is to use this, combined with modern machine learning techniques for time series data, to distinguish between the two groups so reliably that one would be willing to ban a presumed cheater, a decision that can potentially end their career, on the basis of that test alone.
Student Requirements:	Strong programming ability. No prior experience with machine learning is required, although it will allow the project to get further during the summer.
Programs:	This AO can be done under the following programs:   Program    Available To        SURF    Caltech students only  Click on a program name for program info and application requirements.