Different habitat suitability models yield different least-cost path distances for landscape genetic analysis

摘要: Abstract Habitat suitability models (HSMs) are used to describe and predict species distributions based on multiple ecological variables occurrence data. HSMs may also provide a probabilistic identification of least-cost path (LCP) distances in landscape genetics. However, while several studies for these purposes, the performance different genetic analysis and, therefore, consequences model choice have not been carefully explored. In this study, we large dataset wolf genotypes (Canis lupus; n = 923) that were non-invasively sampled central northern Italian Apennines western Alps, aiming (i) estimate LCP derived from ten (ii) quantify correlation between inter-individual using three statistical procedures: partial Mantel tests, regression distance matrices (MRDM) linear mixed effect models. All explained better than Euclidean distances, irrespective applied test. by significantly (paired t-test, P ≤ 0.0001), especially “flexible discriminant analysis” (FDA) “boosted trees” (BRT) “factorial decomposition Mahalanobis distances” (MADIFA) MRDM showed highest coefficient (β) with indicating strong LCPs distances. Results our case study suggest should be compared model-choice procedures identify best fitting HSM analysis.