Testing the accuracy of 3D automatic landmarking via genome-wide association studies

Abstract

BackgroundVarious advances in 3D automatic phenotyping and particularly in landmark-based geometric morphometric methods have been made, but only a few studies have tested the reliability of such automatic procedures in morphometric analyses. It is generally accepted that automatic landmarking compromises the capture of the actual biological variation, and this not only affects its performance to effectively detect differences among sample means but also the structure of covariance matrices. However, no studies have directly tested the actual impact of such landmarking approaches in analyses requiring a large number of specimens and for which the precision of phenotyping is crucial to capture an actual biological signal adequately. ResultsHere, we use a recently developed 3D atlas-based automatic landmarking method to test its accuracy in detecting QTLs associated with craniofacial development of the house mouse skull and lower jaws for a large number of specimens (circa 700) that were previously phenotyped via a semiautomatic landmarking method complemented with manual adjustment. We compare both landmarking methods with univariate and multivariate mapping of the skull and the lower jaws. In the univariate mapping, the automatic approach failed to recover the same SNPs and found only 1 out of 17 previously identified QTLs for the skull, but found one new QTL. Similarly, for the lower jaws, the automatic approach failed to recover the same SNPs but found 2 neighbouring SNPs for 1 out of 8 previously identified QTLs. For centroid size, the same general results were recovered by the automatic method for both the skull and lower jaws, with the same peak SNP being found for the lower jaws. In the multivariate mapping, the automatic approach did not detect the same markers nor QTLs having their regions overlapping with the ones identified with the semi-automatic procedure for the skull, while the same marker, which is also the peak SNP and sole QTL, was recovered by the automatic pipeline for lower jaws. ConclusionOur results confirm the notion that information is lost in the automated landmarking procedure but somewhat dependent on the analyzed structure. The automatic method seems to capture certain types of structures slightly better, such as lower jaws whose shape is almost entirely summarized by its outline and could be assimilated as a 2D flat object. By contrast, the more apparent 3D features exhibited by a structure such as the skull are not adequately captured by the automatic method. We conclude that using 3D atlas-based automatic landmarking methods requires careful consideration of the experimental question and the cautious interpretation of their results.