Danielle Braun
Senior Research Scientist, Department of Biostatistics, Harvard T.H. Chan School of Public Health
The BayesMendel Lab at Dana-Farber Cancer Institute is pioneering the use of advanced statistical modeling, machine learning, and genetic science to develop tools that help identify individuals with inherited cancer-causing mutations and estimate their risk of developing certain cancers. These efforts are co-led by Giovanni Parmigiani, PhD, Professor in the Department of Data Science at Dana-Farber and Danielle Braun, Senior Research Scientist in the Department of Biostatistics at the Harvard T.H. Chan School of Public Health. Their work is not only shaping care within Dana-Farber but is also being widely adopted by research and clinical institutions across the country.
Providing Personalized Cancer Risk Estimates
The Lab’s innovative tools integrate Mendelian genetics (how diseases are inherited), Bayesian probability theory (a method for updating predictions based on new data), and mutation-specific information. By combining a patient’s family history, genetic test results, and findings from the scientific literature, the Lab produces personalized, probabilistic cancer risk estimates.
“Our models calculate lifetime cancer incidence within families and account for the variable penetrance of mutations,” Parmigiani explains. This means the models consider how cancer has affected different family members and how likely it is that a specific genetic mutation will lead to cancer—since not all mutations cause disease in everyone who has them.
“Advanced Bayesian statistical methods then synthesize this data into accurate and actionable predictions,” he says, referring to the Lab’s use of sophisticated algorithms to turn complex data into clear, useful insights.
The predictions generated are more than theoretical—they directly impact patient care.
Meeting a Pressing Need
“The underlying objective of these tools is to identify individuals most likely to benefit from genetic testing and tailored cancer screening, especially when testing resources are limited or reimbursement is constrained.”
Giovanni Parmigiani, PhD
While genetic testing has become more accurate and affordable, challenges remain in ensuring equitable access and interpreting results—especially when they fall into uncertain or ambiguous categories. The BayesMendel Lab’s risk models help navigate these complexities by contextualizing test outcomes within broader family and genetic data.
“The underlying objective of these tools is to identify individuals most likely to benefit from genetic testing and tailored cancer screening, especially when testing resources are limited or reimbursement is constrained,” Parmigiani says. identifying individuals at higher risk, clinicians can recommend earlier screenings, preventive treatments, or lifestyle changes that may reduce cancer risk. For patients, this means more personalized care and potentially better outcomes.
Putting the Tools into Action
Developed with support from the National Institutes of Health (NIH) and other federal research grants, these tools are not only used within Dana-Farber but have been adopted by a growing network of clinicians and genetic counselors nationwide. They’ve been validated and embedded into clinical decision support platforms such as CancerGene, CRA Health, and CancerIQ—systems that help providers interpret genetic data and guide patient care.
“CancerGene alone has over 4,000 users in more than 75 countries, and CRA Health is accessed by more than 15,000 users each month,” Parmigiani notes, underscoring the global reach and practical utility of the Lab’s work.
By embedding these models into the tools clinicians already use, the BayesMendel team has ensured that risk prediction is not a separate process—it’s part of the everyday workflow. This seamless integration helps providers identify high-risk individuals earlier and more accurately, enabling timely interventions that can change the course of care.
Raising Awareness of the Need for Predictive Tools
Parmigiani and his team have shared their research broadly to raise awareness of the importance of predictive tools in cancer care. For instance, they published a paper in eLife in 2021 that looks at the training and functionality utilized by PanelPRO (now renamed Fam3PRO), a novel multi-syndrome cancer risk prediction model that Parmigiani and Braun developed. They also published a study in Genetic Epidemiology in 2022 that looks at the statistical framework and algorithms underlying Fam3PRO, which incorporates updated genetic associations and risk factors, reflecting the latest scientific discoveries in hereditary cancer syndromes. Most recently, a 2023 publication in Cancers explored the value of MyLynch, a tool developed by the BayesMendel lab that helps clinicians and individuals assess risk of a hereditary syndrome that increases risk of colorectal and other cancers by tailoring risk assessments and interventions to individual genetic profiles, reinforcing the importance of patient-facing resources in precision medicine.
He adds that this is particularly important in hereditary syndromes, where early action can significantly alter outcomes. It can also be critical for underserved populations, where access to genetic counseling and testing may be limited.
Looking Ahead
Looking to the future, Parmigiani’s team is refining their models using large-scale genomic data, machine learning techniques, and advanced statistical methods to improve accuracy, generalizability, and adaptability to diverse populations. They are also exploring ways to incorporate real-time data from electronic health records and patient-reported outcomes to further personalize risk assessments.
“As these tools become more integrated into clinical care, they hold the potential to make personalized cancer risk assessment more accurate, accessible, and impactful across populations,” Parmigiani says.
Team Members
Giovanni Parmigiani, PhD
Professor in the Department of Data Science at Dana-Farber
