Identifying Personal DNA Methylation Profiles by Genotype Inference

Explore a conference talk from the 2017 IEEE Symposium on Security & Privacy that delves into the privacy implications of DNA methylation data. Learn how releasing this type of biomedical information can lead to privacy issues similar to releasing one's actual genome. Discover the process of inferring parts of someone's genome using a small subset of methylation regions influenced by genomic variants, and understand the high accuracy of re-identification techniques. Examine the proposed cryptographic scheme for privately classifying tumors, which combines random forests and homomorphic encryption to enable privacy-respecting medical diagnoses in clinical settings. Gain insights into the performance and applications of this novel approach to protecting sensitive biomedical data.