Statistical considerations for preclinical studies