English: Principal components analysis is a form of dimension reduction which simplifies substructure in multi-variant data. With genetic data this often leads to clear separation if samples come from populations that are genetic distant enough from each other. However, when used on samples within a population, one typically doesn't see a sharp distinction. When applied to regions inside inversions, however, this changes. Inversions are regions of the genome where a section as been knocked out and reintegrated backwards. Genetic recombination in those areas is impaired with recombination taking place normally between two copies of the region in the same orientation but impaired with the opposed orientation. This leads to these inversion variants to evolve separately as though they belonged to two populations with no gene flow. The result in a PCA is that a discrete clustering takes place where individuals with two copies of one version are set at either extremes with individuals with one copy of each in the centre. Typically PCs will show something of a continuum. In this example samples of 'Settled' Irish (Green) and Irish Travellers (Pink) are shown. Since the region used was completely inside the inversion, the other two PCs (Eigenvectors 2 and 3) are essentially random with no defined clustering.