Consanguinity in Centre d'Étude du Polymorphisme Humain (CEPH) pedigrees.

A set of Centre d'Étude du Polymorphisme Humain (CEPH) cell lines serves as a large reference collection that has been widely used as a benchmark for allele frequencies in the analysis of genetic variants, to create linkage maps of the human genome, to study the genetics of gene expression, to provide samples to the HapMap and 1000 Genomes projects, and for a variety of other applications. An explicit feature of the CEPH collection is that these multigenerational families represent reference panels of known relatedness, consisting mostly of three-generation pedigrees with large sibships, two parents, and grandparents. We applied identity-by-state (IBS) and identity-by-descent (IBD) methods to high-density genotype data from 186 CEPH individuals in 13 families. We identified unexpected relatedness between nominally unrelated grandparents both within and between pedigrees. For one pair, the estimated Cotterman coefficient of relatedness k1 exceeded 0.2, consistent with one-eighth sharing (eg, first-cousins). Unexpectedly, significant IBD2 values were discovered in both second-degree and parent-child relationships. These were accompanied by regions of homozygosity in the offspring, which corresponded to blocks lacking IBS0 in purportedly unrelated parents, consistent with inbreeding. Our findings support and extend a 1999 report, based on the use of short tandem-repeat polymorphisms, that several CEPH families had regions of homozygosity consistent with autozygosity. We benchmarked our IBD approach (called kcoeff) against both RELPAIR and PREST software packages. Our findings may affect the interpretation of previous studies and the design of future studies that rely on the CEPH resource.

Inference of relationships in population data using identity-by-descent and identity-by-state.

It is an assumption of large, population-based datasets that samples are annotated accurately whether they correspond to known relationships or unrelated individuals. These annotations are key for a broad range of genetics applications. While many methods are available to assess relatedness that involve estimates of identity-by-descent (IBD) and/or identity-by-state (IBS) allele-sharing proportions, we developed a novel approach that estimates IBD0, 1, and 2 based on observed IBS within windows. When combined with genome-wide IBS information, it provides an intuitive and practical graphical approach with the capacity to analyze datasets with thousands of samples without prior information about relatedness between individuals or haplotypes. We applied the method to a commonly used Human Variation Panel consisting of 400 nominally unrelated individuals. Surprisingly, we identified identical, parent-child, and full-sibling relationships and reconstructed pedigrees. In two instances non-sibling pairs of individuals in these pedigrees had unexpected IBD2 levels, as well as multiple regions of homozygosity, implying inbreeding. This combined method allowed us to distinguish related individuals from those having atypical heterozygosity rates and determine which individuals were outliers with respect to their designated population. Additionally, it becomes increasingly difficult to identify distant relatedness using genome-wide IBS methods alone. However, our IBD method further identified distant relatedness between individuals within populations, supported by the presence of megabase-scale regions lacking IBS0 across individual chromosomes. We benchmarked our approach against the hidden Markov model of a leading software package (PLINK), showing improved calling of distantly related individuals, and we validated it using a known pedigree from a clinical study. The application of this approach could improve genome-wide association, linkage, heterozygosity, and other population genomics studies that rely on SNP genotype data.

Thalamus segmentation from diffusion tensor magnetic resonance imaging.

In this paper, we propose a semi-automatic thalamus and thalamus nuclei segmentation algorithm from diffusion tensor magnetic resonance imaging (DT-MRI) based on the mean-shift algorithm. Comparing with existing thalamus segmentation algorithms which are mainly based on K-means algorithm, our mean-shift based algorithm is more flexible and adaptive. It does not assume a Gaussian distribution or a fixed number of clusters. Furthermore, the single parameter in the mean-shift based algorithm supports hierarchical clustering naturally.

Brain structure segmentation from MRI by geometric surface flow.

We present a method for semiautomatic segmentation of brain structures such as thalamus from MRI images based on the concept of geometric surface flow. Given an MRI image, the user can interactively initialize a seed model within region of interest. The model will then start to evolve by incorporating both boundary and region information following the principle of variational analysis. The deformation will stop when an equilibrium state is achieved. To overcome the low contrast of the original image data, a nonparametric kernel-based method is applied to simultaneously update the interior probability distribution during the model evolution. Our experiments on both 2D and 3D image data demonstrate that the new method is robust to image noise and inhomogeneity and will not leak from spurious edge gaps.