Noise peaks will never be fragmented as we loop after chemicals when generating scans, and there are no chemicals associated with noise peaks. See https://github.com/glasgowcompbio/vimms/blob/master/vimms/MassSpec.py#L623-L630. Not sure if this is the best thing to do ..