How big are the interference peaks? Could you try to identify these peaks via MSMS instead of using the selective MRM?