Machine learning model to predict risk assessment of a child inheriting a genetic disorder

Advancements in Machine Learning (ML) have revolutionised precision medicine, particularly in predicting and preventing genetic disorders. This study provides a comprehensive analysis of ML models designed for the risk assessment of genetic disorders in children, based on clinical history and pedigree analysis. Leveraging a diverse dataset of familial genetic profiles, clinical outcomes, and medical histories, we employed ML algorithms to identify inheritance patterns, assess genetic risks for Mendelian disorders, and predict disease recurrence. The analysis incorporates several ML algorithms, including Gradient Boosting, XGBoost, Random Forest, Logistic Regression, Naive Bayes, and Support Vector Machines (SVM), with the Gradient Boosting model achieving the highest mean cross-validation score of over 0.99. Designed for primary care physicians and healthcare professionals, this model aids in genetic counselling by predicting genetic disorder recurrence based on family history. The paper also addresses ethical and legal considerations, emphasising the importance of genetic counselling and informed decision-making. This tool is intended to support, not replace, medical professionals. This work advances personalised risk assessment for Mendelian single-gene disorders, including autosomal dominant, autosomal recessive, and X-linked recessive disorders, contributing to the field of genomic medicine and facilitating effective family planning strategies. Author summaryGenetic disorders are posing significant socio-economic, health, and psychological burdens due to the risk of inheritance, which can impact future generations with health complications. Genetic counselling has emerged as an important tool to help manage this issue, offering guidance to susceptible families. Recent advancements in Artificial Intelligence (AI) and Machine Learning (ML) have introduced new avenues for predicting the risk of genetic disorders with greater accuracy. The aim was to leverage machine learning models that would predict the risk assessment for single gene chromosomal Mendelian disorders. To train these models, patient data was collected from a clinical geneticist at a renowned hospital, all of whom had confirmed genetic testing results. This data was used to train the models that predicted the likelihood of the next child inheriting a genetic disorder. The initial results are promising, demonstrating a good fit with the data. However, it should be noted that larger sample sizes are needed to improve the accuracy of any model. With more extensive data, its predictive capabilities can be significantly enhanced. This tool has the potential to be a resource for genetic counsellors and primary healthcare physicians, aiding them in providing more accurate risk assessments and personalised guidance to families.