Primary Ovarian Insufficiency (POI) is a heterogeneous group of disorders with an incidence of 1:10,000 women by age 20, 1:1,000 by age 30, 1:100 by age 40. POI describes the progression toward the cessation of ovarian function and can occur in the most serious form as primary amenorrhea (PA), with absent pubertal development and/or ovarian dysgenesis (OD), or in milder phenotype with post-pubertal onset and secondary amenorrhea (SA). Several are the etiological causes that may induce ovarian dysfunction, among which the genetic component is considered prevalent (as supported by the occurrence of families with more than one affected women and the existence of several idiopathic cases) but highly heterogeneous. 46,XX non-syndromic women showing the most severe phenotype, characterized by the absence of pubertal development with PA and OD, are very rare but the search for genetic variations in this extreme phenotype may be more effective in identifying novel pathogenic mechanisms. Hence, using high resolution array-CGH we searched for rare high-penetrant CNVs involving genes essential for ovarian function in a cohort of 67 46,XX non-syndromic patients affected by PA, namely 53 sporadic (79.1%) and 14 familial (20.9%) cases. 28 out of 67 women resulted positive to array-CGH analysis because having at least one rare ?ovarian? CNV: a total number of 32 CNVs involving 37 ovarian genes was selected. Population from Database of Genomic Variants (DGV) was used to evaluate the rarity of POI CNVs, but it does not match to the ideal set of controls for POI disease (neither age nor gender of DGV controls are known). Thus, to better understand the CNVs contribution in disease onset, the rare ?ovarian? CNVs found in patients according to DGV were searched in an ad hoc control cohort, previously screened by array-CGH, consisting in 140 healthy women with normal reproductive life and physiological menopause. 28 out of 32 rare ?ovarian? CNVs detected in patients were not found in the control group thus supporting their role in the POI?s pathogenesis. Moreover, to evaluate the presence of a significant enrichment in ovarian genes in the POI group, array-CGH of the ad hoc control cohort were analyzed with the same approach adopted for patients cohort and 49 CNVs involving 54 ovarian genes were selected. Several statistical analyses were performed comparing patients? to controls? data and revealed no significant differences. Nevertheless, the CNVs found in the POI cohort containing ovarian genes are more harmful respect to the CNVs identified in the controls cohort. The 37 genes perturbed or possibly perturbed by POI CNVs are implicated in several ovarian processes (e.g., regulation of cytoskeleton dynamics for oocytes asymmetric division, maintenance of oocytes genomic integrity, ovarian differentiation, follicular development, and meiotic resumption), thus supporting their involvement in POI etiology. Validation and characterization of selected CNVs, as well as the study of a possible gene perturbation at mRNA level, was also crucial in order to perform a correct genotype-phenotype correlation and to propose new candidate genes for POI disease (e.g. TP63, VLDLR). 39 out of 67 women resulted negative for rare ?ovarian? CNVs (58.2%) suggesting to combine different genomic approaches in order to increase the detection rate of the disorder. Hence, 17 out of 67 collected patients, were submitted to a preliminary WES analysis searching for rare SNVs in a total of 191 genes selected from array-CGH data, and literature regarding POI and ovary. The WES preliminary analysis allowed to confirm the importance of some array CGH new candidate genes in POI onset (e.g. VLDLR) and the complex heterogeneity of POI. The combination of these molecular evidences, with major or minor contribution, might have been at the basis of POI supporting the existence of a disease genetic model characterized by oligogenic heterozygosity (i.e., the simultaneous presence in a single patient of multiple heterozygous quantitative variants/rare mutations, both de novo and/or inherited, affecting multiple genes). The present approach using both array-CGH and WES techniques, resulted an efficient tool to identify rare variants (CNVs and SNVs) involving both genes already reported in POI, and new candidate genes with a role in oocyte maturation and differentiation. The results of this study are promising to expand the knowledge about the molecular pathways involved in POI pathogenesis and probably provide the basis for a more accurate genetic diagnosis of POI patients.