Up to 40% of genetic and rare disorders (RD) present facial dysmorphologies. Visual assessment of facial gestalt is commonly used for clinical diagnosis, health management and treatment monitoring. Quantitative approaches to facial phenotypes are more objective and provide first diagnoses of RD with relatively high accuracy, but are mainly based on populations of European descent, disregarding the influence of population ancestry. Here we assessed the facial phenotypes associated to four genetic disorders in a Latino-American population from Colombia. We recorded the coordinates of 18 facial landmarks in 2D images from 79 controls 51 pediatric individuals diagnosed with Down (DS), Morquio (MS), Noonan (NS) and Neurofibromatosis type 1 (NF1) syndromes. We quantified facial differences using Euclidean Distance Matrix Analysis (EDMA) and assessed the diagnostic accuracy of Face2gene, an automatic deep learning algorithm with widespread use in the clinical practice.Quantitative comparisons indicated that individuals diagnosed with DS and MS were associated with the most severe phenotypes, with 58.2% and 65.4% of facial traits significantly different as compared to controls. The percentage decreased to 47.7% in NS and to 11.4% in NF1. Each syndrome presented a characteristic pattern of facial dysmorphology, supporting the potential of facial biomarkers for disorder diagnosis. However, our results detected population-specific traits in the Colombian population as compared to the facial gestalt described in literature for DS, NS and NF1. When clinical diagnosis based on genetic testing was used to verify the diagnosis based on 2D facial pictures, our results showed that Face2Gene accuracy was very high in DS, moderate in NS and NF1, and very low in MS, with low gestalt similarity scores in highly admixed individuals. Our study underscores the added value of precise quantitative comparison of facial dysmorphologies in genetic and rare disorders and the need to incorporate populations with diverse contributions of Amerindian, African and European ancestry components to further improve automatic diagnostic methods.