From Data to Dialogue: How AI can Shape the Future of Genetic Counseling
Writen By: Julia Richie
In the age of genomics, genetic counselors play a pivotal role in helping individuals and families understand, navigate and interpret the implications of complex genetic information. Their combined expertise in genetic sciences and psychosocial care places them in the unique role of interpreters and advocates in an increasingly intricate world of genetics and patient care. For genetic counselors, the expansion in the availability of genetic testing presents a double-edged sword: while more data enables more precise diagnoses, it is often also accompanied by novel interpretive challenges.
The completion of the Human Genome Project in 2003 was a turning point. Ushered in by the era of precision medicine, contemporary methods for genetic testing and sequencing, such as whole-genome (WGS) and whole-exome sequencing (WES), can now generate a massive amount of genetic data in mere minutes. The value of genetic counselors takes center stage through their ability to translate this data into personalized insights for patients. However, combing through thousands of genetic variants can be time-consuming, especially in cases of rare diseases, which often involves cross-referencing phenotypes and genetic findings with scientific literature and database platforms. In this landscape, the rate-limiting step to patient care is not the generation of this data, but rather its interpretation. This is where the use of artificial intelligence (AI), particularly those based on machine learning (ML), may offer transformative potential to reduce this burden within the genetic counseling profession, functioning as an efficient assistant rather than a replacement for human judgment.
Among the most promising of these tools is Google AI’s DeepVariant, which uses deep neural networks to call genetic variants by “visualizing” sequencing data and making predictions based on image pattern recognition. In a 2018 study published in Nature Scientific Reports, this platform was tested using WES data from patients suspected of having Mendelian disorders and achieved an F1 score of 0.94 for indel (insertion/deletion) calling. For small nucleotide variants (SNVs), this tool stood out in precision and accuracy with a 0.98 F1 score, making it the top-performing variant caller at the time. DeepVariant, along with similar platforms, can help support genetic counselors in interpreting genetic data by omitting the need for visual inspection using genome browsers while still maintaining expert-level precision.
Another useful tool is AMELIE (Automatic Mendelian Literature Evaluation), an AI tool that helps link patients’ phenotypic symptoms to candidate gene information by analyzing all existing PubMed abstracts. A study published in 2020 by Science Translational Medicine found that the top 11 genes ranked by AMELIE out of 127 candidate genes assisted in the diagnosis of singleton variant patients. In over 90% of cases, prioritized genes were later confirmed to contain the causative gene of the disorders. Manually searching publications to isolate a single gene for an ambiguous phenotype can be a highly laborious and time-consuming process for a genetic counselor, but can be done in a fraction of the time required by an AI tool, without compromising diagnostic accuracy.
Additionally, some AI tools may prove useful in pre-screening and risk assessment environments, particularly through chatbots in clinical settings. A study presented at the American Association for Cancer Research (AACR) evaluated an AI chatbot to perform hereditary cancer risk assessments in OB/GYN clinics across the United States. Following the chatbot’s collection of family histories and evaluation of hereditary cancer risk, the findings proved to be consistent with those of the National Comprehensive Cancer Network (NCCN), where 1 in 4 women met the criteria for genetic testing based on family cancer history and inherited cancer risk. This type of automated tool could help evaluate pre-testing patients and alleviate
AI tools may contribute to alleviating a cognitive burden on genetic counselors, while providing efficient and increasingly accurate results. However, AI is not designed to replace genetic counselors. Unlike many professions in which automated technology has superseded the need for human labour, human connection in genetic counseling is indispensable. Patients often seek genetic counseling in times of vulnerability, where the communication, advocacy and empathy provided by a genetic counselor is a pillar in a positive and patient-centred experience in the seemingly enigmatic world of genetics. Because of this, the role undertaken by AI in the future of genetic counseling may likely resemble a partnership rather than a replacement. Implementing AI in the profession may enable genetic counselors more time to screen new patients for risk assessment, address additional questions and focus their time with patients who have complex results.
Resoruces: