Two lines of evidence are promoted this proposal. First, recently, smaller microdeletions in the region between the human SNRPN and UBE3A genes have been reported in several cases with features consistent with PWS, including childhood obesity, hyperphagia, and hypogonadism. Second, interestingly, recent studies of genomewide survey of imprinting genes in mouse brains have revealed a rather complex and tissue specific imprinting expression profile noncoding RNA transcripts from Snrpn to Ube3a. Many of these transcripts, including the host transcript for SnoRNAs, have paternal expression bias in embryonic brains but bi-allelic expression in adult preoptic area of hypothalamus and medial prefrontal cortex. We hypothesize that imprinting profile from SNRPN to UBE3A in human is different in mice. We further hypothesize that deficiency of both HBII-85 and HBII-52 are responsible for the full spectrum of clinical features in PWS. The difference in imprinting expression between human and mice may contribute to the difficulty of modeling human PWS in previous mouse models. We propose a series of experiments to compare the imprinting expression profile of transcripts from SNRPN to UBE3A region between human and mice using human PWS postmortem brain tissues with a deletion of 15q11-q13 and mice with a paternal deletion from Snrpn to Ube3a available to us. We propose to generate a brain specific deletion of Snrpn to Ube3a in mice to better model human PWS. The observation of bi-allelic expression of many host transcripts in hypothalamus from Snrpn to Ube3a may argue that hypothalamus related obesity and hyperphagia caused by biallelic expression transcripts would be seen in homozygous deletion mice if we could prevent the early lethality in mice carrying a deletion. We have obtained mice with loxP sites inserted into the Snrpn and Ube3a genes but without a deletion between them (S-U-floxed mice). Our pilot study of breeding S-U-floxed and Nestin-Cre and CamkII-cre has confirmed the feasibility of generating mice with brain-specific deletion from Snrpn to Ube3a. Our preliminary data have shown that perinatal lethality is not associated with brain-specific and postnatal deletion from Snrpn to Ube3a. Thus, mice with a brain specific deletion may be better model to study human PWS. Data from this study will allow us to determine if these mice are suitable models for dissecting the pathogenesis of PWS. The objective of the proposed project clearly meets the mission of FPWR.