Researchers Develop Phenome-Risk Classification Machine Learning Algorithm to Identify Patients Who Are at High Risk for Developmental Stuttering; Results Presented at ASHG 2020 Virtual Annual Meeting (October 27-30)

On Day 3 (Thursday, October 29) of the American Society of Human Genetics (ASHG) 2020 Virtual Annual Meeting (, one of the most interesting presentations was on the subject of developmental stuttering. Douglas M. Shaw (a graduate student in the Vanderbilt Genetics institute, Vanderbilt University) gave talk entitled “Applying a Phenome Risk Classifying Model to Identify Undiagnosed Developmental Stuttering Cases in a Biobank for Genome Wide Association Analysis.” In the abstract to his talk, Shaw described “developmental stuttering” as a speech disorder characterized by a disturbance in fluency and speech pattern, with an adult prevalence of 1-3% in the US. Despite twin-based studies showing ~50% heritability, the genetic etiology of stuttering is still largely unknown. No population-based genome wide association analysis (GWAS) has yielded variants that reach genome-wide significance, Shaw and colleagues wrote. Shaw noted that within Vanderbilt’s Electronic Health Record-linked biorepository (BioVU), only 142 cases of stuttering have diagnostic ICD9/10 (ICD9-307.0, ICD10-F98.5, ICD9-315.35, ICD10-F80.81, ICD10-R47.82) codes out of 92,762 genotyped samples, suggesting a large portion of people who stutter are not well-captured within the EHR. To address this case acquisition issue and provide a large enough sample set to power a GWAS, Shaw and colleagues developed a phenome-risk classification machine learning algorithm to identify patients who are at high risk for developmental stuttering.
Login Or Register To Read Full Story