Responsible Conduct of Research Part II
An Overview of Pathogenicity Prediction Methods
DateTuesday, September 18, 2018 - 4:00pm
AbstractProving a causal link between a gene variant and disease is time consuming and expensive. By utilizing models which try to predict which mutations will be harmful, we can help improve quality of research as well as patient care. These models use a variety of methods and data to determine whether given genetic mutations are likely to be associated with disease. We'll be reviewing several current methods for pathogenicity prediction, and discussing the importance of these methods for clinical studies and understanding the genetics of disease.
Sodium Channels Pathologies and Statistical Issues in Pathogenicity Prediction
DateTuesday, September 25, 2018 - 4:00pm
AbstractThis talk will focus on the statistical insides of pathogenicity prediction, classification methodologies, and assessments. Our motivating example will use naive Bayes as a classifier, cross-validation, and receiver operating characteristics to assess its performance. We will then see how this applies to a family of sodium channels proteins, what the mutations in this protein family predicts and some insights into the prediction’s shortcomings.