No abstract

This work details a highly efficient implementation of the 3D scale-invariant feature transform (SIFT) algorithm, for the purpose of machine learning from large sets of volumetric medical image data. The primary operations of the 3D SIFT code are implemented on a graphics processing unit (GPU), including convolution, sub-sampling, and 4D peak detection from scale-space pyramids. The performance improvements are quantified in keypoint detection and image-to-image matching experiments, using 3D MRI human brain volumes of different people. Computationally efficient 3D keypoint descriptors are proposed based on the Binary Robust Independent Elementary Feature (BRIEF) code, including a novel descriptor we call Ranked Robust Independent Elementary Features (RRIEF), and compared to the original 3D SIFT-Rank method(Toews and Wells III, 2013). The GPU implementation affords a speedup of approximately 7X beyond an optimised CPU implementation, where computation time is reduced from 1.4 seconds to 0.2 seconds for 3D volumes of size (145, 174, 145) voxels with approximately 3000 keypoints. Notable speedups include the convolution operation (20X), 4D peak detection (3X), sub-sampling (3X), and difference-of-Gaussian pyramid construction (2X). Efficient descriptors offer a speedup of 2X and a memory savings of 6X compared to standard SIFT-Rank descriptors, at a cost of reduced numbers of keypoint correspondences, revealing a trade-off between computational efficiency and algorithmic performance. The speedups gained by our implementation will allow for a more efficient analysis on larger data sets. Our optimized GPU implementation of the 3D SIFT-Rank extractor is available at (https://github.com/CarluerJB/3D_SIFT_CUDA).

The first Genome Wide Association Studies (GWAS) shed light on the concept of missing heritability. It constitutes a mystery with transcending consequences from plant to human genetics. This mystery lies in the fact that a large proportion of phenotypes are not explained by unique or simple genomic modifications. One has to invoke genetic interactions among different loci, also known as epistasis, to partly account for it. However, current GWAS statistical models are moderately scalable, very sensitive to False Discovery Rate (FDR) corrections and, even combined with High Performance Computing (HPC), they can take years to evaluate for a full combinatorial epistatic space for a single phenotype. Here we propose a modeling approach, named Next-Gen GWAS (NGG) that evaluates, within hours, >60 billions of single nucleotide polymorphism (SNP) combinatorial first-order interactions, on a reasonable computer power. We first benchmark NGG on state of the art GWAS model results, and apply this to Arabidopsis thaliana providing 2D epistatic maps at gene resolution. We demonstrate on several phenotypes that a large proportion of the missing heritability can i) be retrieved with this modeling approach, ii) indeed lies in epistatic interactions and iii) can be used to improve phenotype prediction.