CNVnator has been used successfully in some studies for analyzing whole-genome germline samples.
Also consider a more general structural variant caller like LUMPY. While you're at it, if you try multiple callers you can use MetaSV to derive a consensus set of structural variant calls.
Thanks Etal for your valuable suggestions. I too explored CNVnator as being used in some of studies. Let me see if CNVnator works for me or not.