Infection-triggered familial or recurrent cases of acute necrotizing encephalopathy caused by mutations in a component of the nuclear pore, RANBP2.

Acute necrotizing encephalopathy (ANE) is a rapidly progressive encephalopathy that can occur in otherwise healthy children after common viral infections such as influenza and parainfluenza. Most ANE is sporadic and nonrecurrent (isolated ANE). However, we identified a 7 Mb interval containing a susceptibility locus (ANE1) in a family segregating recurrent ANE as an incompletely penetrant, autosomal-dominant trait. We now report that all affected individuals and obligate carriers in this family are heterozygous for a missense mutation (c.1880C-->T, p.Thr585Met) in the gene encoding the nuclear pore protein Ran Binding Protein 2 (RANBP2). To determine whether this mutation is the susceptibility allele, we screened controls and other patients with ANE who are unrelated to the index family. Patients from 9 of 15 additional kindreds with familial or recurrent ANE had the identical mutation. It arose de novo in two families and independently in several other families. Two other patients with familial ANE had different RANBP2 missense mutations that altered conserved residues. None of the three RANBP2 missense mutations were found in 19 patients with isolated ANE or in unaffected controls. We conclude that missense mutations in RANBP2 are susceptibility alleles for familial and recurrent cases of ANE.

Human PAML browser: a database of positive selection on human genes using phylogenetic methods.

With the recent increase in the number of mammalian genomes being sequenced, large-scale genome scans for human-specific positive selection are now possible. Selection can be inferred through phylogenetic analysis by comparing the rates of silent and replacement substitution between related species. Maximum-likelihood (ML) analysis of codon substitution models can be used to identify genes with an accelerated pattern of amino acid substitution on a particular lineage. However, the ML methods are computationally intensive and awkward to configure. We have created a database that contains the results of tests for positive selection along the human lineage in 13,721 genes with orthologs in the UCSC multispecies genome alignments. The Human PAML Browser is a resource through which researchers can search for a gene of interest or groups of genes by Gene Ontology category, and obtain coding sequence alignments for the gene and as well as results from tests of positive selection from the software package Phylogenetic Analysis by Maximum Likelihood. The Human PAML Browser is available at http://mendel.gene.cwru.edu/adamslab/pbrowser.py.

An empirical test for branch-specific positive selection.

The use of phylogenetic analysis to predict positive selection specific to human genes is complicated by the very close evolutionary relationship with our nearest extant primate relatives, chimpanzees. To assess the power and limitations inherent in use of maximum-likelihood (ML) analysis of codon substitution patterns in such recently diverged species, a series of simulations was performed to assess the impact of several parameters of the evolutionary model on prediction of human-specific positive selection, including branch length and d(N)/d(S) ratio. Parameters were varied across a range of values observed in alignments of 175 transcription factor (TF) genes that were sequenced in 12 primate species. The ML method largely lacks the power to detect positive selection that has occurred since the most recent common ancestor between humans and chimpanzees. An alternative null model was developed on the basis of gene-specific evaluation of the empirical distribution of ML results, using simulated neutrally evolving sequences. This empirical test provides greater sensitivity to detect lineage-specific positive selection in the context of recent evolutionary divergence.

A catalog of nonsynonymous polymorphism on mouse chromosome 16.

Numerous phenotypic traits differ among inbred mice, and the genetic diversity of inbred strains has been exploited in studies of quantitative trait loci (QTL). Sequencing the mouse genome has resulted in improved tools for the study of QTL, but a comprehensive catalog of sequence variants between strains would be of great value in identifying and testing potentially causative alleles. A/J DNA was included in the Celera shotgun sequence of the mouse genome and C57BL/6 DNA was sequenced by an international consortium. We have resequenced A/J and B6 DNA to cover nearly all of the protein-coding portions of mouse Chromosome 16, revealing that there are 106 nonsynonymous substitutions in 74 of the 779 genes on the chromosome. The pattern of substitution is more similar to the spectrum of benign polymorphism in the human population than it is to human disease-causing mutations. In mouse, polymorphic variants tend to be associated with one another on large haplotypes; this pattern also holds true for nonsynonymous polymorphism. However, sufficient fragmentation of haplotypes is present to suggest that only a very-high-resolution haplotype map will enable effective inference of alleles in additional strains.