SUMMARY: Whole-exome sequencing (WES) has extensively been used in cancer genome studies; however, the use of WES data in the study of loss of heterozygosity or more generally allelic imbalance (AI) has so far been very limited, which highlights the need for user-friendly and flexible software that can handle low-quality datasets. We have developed a statistical approach, ExomeAI, for the detection of recurrent AI events using WES datasets, specifically where matched normal samples are not available. AVAILABILITY: ExomeAI is a web-based application, publicly available at: http://genomequebec.mcgill.ca/exomeai. CONTACT: [email protected] or [email protected] SUPPLEMENTARY INFORMATION: Supplementary data are available at Bioinformatics online.