BackgroundMany chemotherapy agents act by inducing DNA damage, leading to cell death. Resistance and treatment failure limits the efficacy of these agents, and repair of the DNA damage is the commonest cause of resistance. By combining DNA immunoprecipitation with DNA microarray technology, we have developed a novel method to detect, quantify, and localise DNA damage at high resolution throughout the human genome. MethodsWe developed this assay initially in the model yeast organism Saccharomyces cerevisiae, before further development in human dermal fibroblast cells. Cellular DNA was treated in vivo or naked DNA treated in vitro with cisplatin, oxaliplatin, or ultraviolet (UV) irradiation. Using antibodies against cisplatin-modified DNA or cyclobutane pyrimidine dimers, we captured damaged DNA fragments through DNA immunoprecipitation. DNA samples were then amplified, labelled and hybridised to custom-designed DNA microarrays to generate high-resolution genome-wide profiles for DNA damage. FindingsGenome-wide cisplatin, oxaliplatin, and UV-induced damage profiles were generated in human cells and were consistent, reproducible, and correlated well with mathematical models predicting sites of damage induction developed in our laboratory (r=0·7–0·8). Using standard chromatin immunoprecipitation-DNA microarray experiments, we generated genome-wide epigenetic profiles showing increased H3K14 acetylation after platinum exposure. By applying novel bioinformatic outlier detection methods, we identified genomic regions where differences occured between in-vitro and in-vivo damage profiles, and we demonstrated how this information could be integrated and correlated with epigenetic changes. InterpretationWe validated a novel technique using DNA microarrays to sensitively measure DNA damage and repair at high resolution throughout the human genome, and we used the platinum analogue chemotherapy agents and UV irradiation as a model for damage induction. This technology can be applied to other DNA damaging agents and to clinical samples taken from patients after treatment. We believe that by applying this assay, and using our bioinformatic approaches for integrating genomic and epigenetic datasets, we will be able to detect functional biomarkers and genetic signatures predictive of response or toxicity, allowing us to optimise and stratify the use of current therapies. We are applying these assays to clinical samples. FundingUK Medical Research Council, Velindre NHS Trust Charitable Funds.