Eliminating accidental deviations to minimize generalization error and maximize replicability: Applications in connectomics and genomics

Replicability, the ability to replicate scientific findings, is a prerequisite for scientific discovery and clinical utility. Troublingly, we are in the midst of a replicability crisis. A key to replicability is that multiple measurements of the same item (e.g., experimental sample or clinical participant) under fixed experimental constraints are relatively similar to one another. Thus, statistics that quantify the relative contributions of accidental deviations—such as measurement error—as compared to systematic deviations—such as individual differences—are critical. We demonstrate that existing replicability statistics, such as intra-class correlation coefficient and fingerprinting, fail to adequately differentiate between accidental and systematic deviations in very simple settings. We therefore propose a novel statistic, discriminability, which quantifies the degree to which an individual’s samples are relatively similar to one another, without restricting the data to be univariate, Gaussian, or even Euclidean. Using this statistic, we introduce the possibility of optimizing experimental design via increasing discriminability and prove that optimizing discriminability improves performance bounds in subsequent inference tasks. In extensive simulated and real datasets (focusing on brain imaging and demonstrating on genomics), only optimizing data discriminability improves performance on all subsequent inference tasks for each dataset. We therefore suggest that designing experiments and analyses to optimize discriminability may be a crucial step in solving the replicability crisis, and more generally, mitigating accidental measurement error.     

Analysis of the coding regions of <Emphasis Type="Italic">VEGFR3</Emphasis> and <Emphasis Type="Italic">VEGFC</Emphasis> in Milroy disease and other primary lymphoedemas

Milroy disease (hereditary lymphoedema type I, MIM 153100) is a congenital onset primary lymphoedema with autosomal dominant inheritance. Mutations in the gene, vascular endothelial growth factor receptor 3, VEGFR3 (FLT4), are known to cause Milroy disease, but there is uncertainty about the prevalence of VEGFR3 mutations in patients with primary lymphoedema and more specifically in those with a phenotype that resembles Milroy disease. This study aims to address this issue and thereby delineate the Milroy disease phenotype. Fifty-two patients with primary lymphoedema were analysed for mutations in the coding regions of VEGFR3. Patients were divided into four groups: Typical Milroy disease with family history (group I), typical Milroy disease with no family history (group II), atypical Milroy disease (group III), and complex primary lymphoedema (group IV). Results demonstrated that with rigorous phenotyping the likelihood of detecting VEGFR3 mutations is optimised. Mutation prevalence is 75% in typical Milroy patients with a family history (group I) and 68% if positive family history is not a diagnostic criterion. A positive family history is not essential in Milroy disease. The likelihood of detecting VEGFR3 mutations in patients who have a phenotype which is not typical of Milroy disease is very small (<5%). For the 22 mutation positive patients, 14 novel VEGFR3 mutations were identified, two of which were in exon 22 and one in exon 17, confirming that these exons should be included in VEGFR3 analysis. No mutations were found outside the kinase domains, showing that analysis of this part of the gene is not useful for Milroy disease patients. VEGFC, which encodes the ligand for VEGFR3, was sequenced in all patients with typical Milroy disease (groups I and II) and no mutations were identified. F. C. Connell and P. Ostergaard contributed equally to this work. An erratum to this article can be found at     

An altered pattern of circulating apolipoprotein E3 isoforms is implicated in preeclampsia

Preeclampsia is a common pregnancy complication that is an important cause of preterm birth and fetal growth restriction. Because there is no diagnostic test yet available for preeclampsia, we used a proteomic approach to identify novel serum/plasma biomarkers for this condition. We conducted case control studies comparing nulliparous women who developed preeclampsia at 36-38 weeks of gestation with healthy nulliparous women matched by gestational age at sampling. Serum/plasma was depleted of six abundant proteins and analyzed by two-dimensional gel electrophoresis (n = 12 per group) and difference gel electrophoresis (n = 12 per group). Differences in abundance of protein spots were detected by univariate and multivariate statistical analyses. Proteins were identified by mass spectrometry and expression of selected proteins was validated by immunoblotting. Proteins whose concentrations were selectively associated with preeclampsia included apolipoprotein E (apoE), apoC-II, complement factor C3c, fibrinogen, transthyretin, and complement factor H-related protein 2. An increase in a deglycosylated isoform of apoE3 and concomitantly decreased amounts of one apoE3 glycoisoform were identified in preeclamptic plasma and confirmed by immunoblotting. Altered production of these preeclampsia-related apoE3 isoforms might impair reverse cholesterol transport, contributing to arterial damage. These findings point to a novel mechanistic link between preeclampsia and subsequent cardiovascular disease.     

Somatic point mutations in mtDNA control region are influenced by genetic background and associated with healthy aging: a GEHA study

Tissue specific somatic mutations occurring in the mtDNA control region have been proposed to provide a survival advantage. Data on twins and on relatives of long-lived subjects suggested that the occurrence/accumulation of these mutations may be genetically influenced. To further investigate control region somatic heteroplasmy in the elderly, we analyzed the segment surrounding the nt 150 position (previously reported as specific of Leukocytes) in various types of leukocytes obtained from 195 ultra-nonagenarians sib-pairs of Italian or Finnish origin collected in the frame of the GEHA Project. We found a significant correlation of the mtDNA control region heteroplasmy between sibs, confirming a genetic influence on this phenomenon. Furthermore, many subjects showed heteroplasmy due to mutations different from the C150T transition. In these cases heteroplasmy was correlated within sibpairs in Finnish and northern Italian samples, but not in southern Italians. This suggested that the genetic contribution to control region mutations may be population specific. Finally, we observed a possible correlation between heteroplasmy and Hand Grip strength, one of the best markers of physical performance and of mortality risk in the elderly. Our study provides new evidence on the relevance of mtDNA somatic mutations in aging and longevity and confirms that the occurrence of specific point mutations in the mtDNA control region may represent a strategy for the age-related remodelling of organismal functions.     

Design and Characterization of a 52K SNP Chip for Goats

The success of Genome Wide Association Studies in the discovery of sequence variation linked to complex traits in humans has increased interest in high throughput SNP genotyping assays in livestock species. Primary goals are QTL detection and genomic selection. The purpose here was design of a 50–60,000 SNP chip for goats. The success of a moderate density SNP assay depends on reliable bioinformatic SNP detection procedures, the technological success rate of the SNP design, even spacing of SNPs on the genome and selection of Minor Allele Frequencies (MAF) suitable to use in diverse breeds. Through the federation of three SNP discovery projects consolidated as the International Goat Genome Consortium, we have identified approximately twelve million high quality SNP variants in the goat genome stored in a database together with their biological and technical characteristics. These SNPs were identified within and between six breeds (meat, milk and mixed): Alpine, Boer, Creole, Katjang, Saanen and Savanna, comprising a total of 97 animals. Whole genome and Reduced Representation Library sequences were aligned on >10 kb scaffolds of the de novo goat genome assembly. The 60,000 selected SNPs, evenly spaced on the goat genome, were submitted for oligo manufacturing (Illumina, Inc) and published in dbSNP along with flanking sequences and map position on goat assemblies (i.e. scaffolds and pseudo-chromosomes), sheep genome V2 and cattle UMD3.1 assembly. Ten breeds were then used to validate the SNP content and 52,295 loci could be successfully genotyped and used to generate a final cluster file. The combined strategy of using mainly whole genome Next Generation Sequencing and mapping on a contig genome assembly, complemented with Illumina design tools proved to be efficient in producing this GoatSNP50 chip. Advances in use of molecular markers are expected to accelerate goat genomic studies in coming years.     

Selection Signatures in Worldwide Sheep Populations

The diversity of populations in domestic species offers great opportunities to study genome response to selection. The recently published Sheep HapMap dataset is a great example of characterization of the world wide genetic diversity in sheep. In this study, we re-analyzed the Sheep HapMap dataset to identify selection signatures in worldwide sheep populations. Compared to previous analyses, we made use of statistical methods that (i) take account of the hierarchical structure of sheep populations, (ii) make use of linkage disequilibrium information and (iii) focus specifically on either recent or older selection signatures. We show that this allows pinpointing several new selection signatures in the sheep genome and distinguishing those related to modern breeding objectives and to earlier post-domestication constraints. The newly identified regions, together with the ones previously identified, reveal the extensive genome response to selection on morphology, color and adaptation to new environments.     

Phylogenetic tree of vascular plants reveals the origins of aquatic angiosperms

Although aquatic plants are discussed as a unified biological group, they are phylogenetically well dispersed across the angiosperms. In this study, we annotated the aquatic taxa on the tree of vascular plants, and extracted the topology of these aquatic lineages to construct the tree of aquatic angiosperms. We also reconstructed the ancestral areas of aquatic families. We found that aquatic angiosperms could be divided into two different categories: the four aquatic orders and the aquatic taxa in terrestrial orders. Aquatic lineages evolved early in the radiation of angiosperms, both in the orders Nymphaeales and Ceratophyllales and among basal monocots (Acorales and Alismatales). These aquatic orders do not have any extant terrestrial relatives. They originated from aquatic habitats during the Early Cretaceous. Asia would have been one of the centers for early diversification of aquatic angiosperms. The aquatic families within terrestrial orders may originate from other areas besides Asia, such as America or Australia. The lineages leading to extant angiosperms diversified early in underexploited freshwater habitats. The four extant aquatic orders were relicts of an early radiation of angiosperm in aquatic environments. Their extinct ancestors might be aquatic early angiosperms.