Approved Research

Deconvolution of factors that interfere with accurate biomarker interpretation

Lay summary

Precision medicine is based on predicting, preventing and treating diseases in an individualized manner. For example, patients may have similar symptoms but need very different treatment. Some diseases are hard to follow up with treatment responses with clinical exams alone. To help physicians in these cases, a lot of effort has been spent on identification of blood markers that could predict disease diagnosis, disease progression and response to treatment. If successful, these blood markers make treating patients much easier and effective. However, establishing these blood markers for many diseases have been challenging. One of the main issues is identifying blood markers that replicate across multiple studies and centers. Our group hypothesizes that one reason for that is that "normal life" affects the blood levels of these biomarkers. For example: behaviors (nutrition choices, supplements, exercise routine, sleep hygiene); biological cycles (menstruation, circadian rhythm); and acute variances on day of draw (time of the day, last meal, time since wake), among others. There is an emerging body of scientific evidence supporting our hypothesis: (1) Multivitamin targeting hair, skin and nails have high concentrations of biotin, which is known to directly interfere with tests for markers of cardiovascular risk, hormone tests, and tests detecting infectious diseases; (2) Glucose levels are affected by circadian clock (the internal process that regulates the sleep-wake cycle of our bodies); (3) Menstrual cycle may affect cortisol levels, the classic stress response hormone; (4) Sedentary individuals with normal glucose and triglycerides may still be at risk for metabolic diseases, as indicated by an increase in triglyceride levels post-meal.

To investigate whether daily variances may indeed affect biomarkers levels, we will explore the correlation between individuals' characteristics and habits with blood markers. Once we understand which factors affect these blood markers the most, we will develop methods to  correct for these variances. We predict these corrections will lead to more accurate and stable markers.

The project should last about 2 years:

1- The first draft of correlations: 3 months

2- Deeper understanding of pre-selected correlations : 3 months

3- Algorithm to account for top 10 correlations: 6 months

4- Investigate accuracy of corrections on new data: 6 months

5- Reiterate steps 2/3/4 as needed : 6 months