Achieving k-anonymity in DataMarts used for gene expressions exploitation

Abstract

Gene expression profiling is a sophisticated method to discover differences in activation patterns of genes between different patient collectives. By reasonably defining patient groups from a medical point of view, subsequent gene expression analysis may reveal disease-related gene expression patterns that are applicable for tumor markers and pharmacological target identification. When releasing patient-specific data for medical studies privacy protection has to be guaranteed for ethical and legal reasons. k-anonymisation may be used to generate a sufficient number of k data twins in order to ensure that sensitive data used in analyses is protected from being linked to individuals. We use an adapted concept of k-anonymity for distributed data sources and include various customisation parameters in the anonymisation process to guarantee that the transformed data is still applicable for further processing. We present a real-world medical-relevant use case and show how the related data is materialised, anonymised, and released in a data mart for testing the related hypotheses.