Many sporadic and severe pediatric diseases remain undiagnosed despite extensive medical and genetic test evaluations. Diagnosis is crucial for prognosis, specific and timely treatment, family planning, and to understand the pathomechanisms of diseases. We created the Telethon Undiagnosed Diseases Program (TUDP) to provide a systematic approach to improve diagnostic yield and provide an overview of the underlying genetic mechanisms. All cases were submitted together with electronic phenotyping from approximately 50 clinical geneticists from pediatric genetics and clinic centers or via an online platform. Procedures were standardized across centers through periodic online meetings to achieve more accurate and reproducible phenotyping. Eligibility criteria were based on the severity, complexity, and negative results of full genetic analysis. Genomic analyses were all performed at least in trios and the case reanalysis and matching process was periodically repeated and targeted for all still negative cases. Over five years, TUDP studied 1,098 families providing 778 full clinical genetic reports. A conclusive diagnosis was provided in 49.1% of cases, with the identification of mutations in over 200 different genes, while a further number of mutations in new genes are being matched to find second cases. The general overview indicates that the vast majority (71%) of the causative variants were de novo, either with autosomal dominant (68%) or X-linked (3%) alleles. Recessive forms (either autosomal or X-linked) explained the remaining 29% of cases, with homozygous mutations identified in only 10% of cases. The high percentage of de novo mutations may reflect the postponement of parenthood and thus the number of new cases is expected to increase in the coming years. This may have important implication for genetic counselling and social policy. We conclude that a systematic approach can solve about 50% of missing diagnosis due the heterogeneity of genetic causes, while the remaining cases may be due to unique and therefore elusive mechanisms that require extended matching. The multicenter TUDP model should be transferred to clinical settings to avoid diagnostic delays or misdiagnoses in children.