Accepted Posters
Attention Conference Presenters - please review the Speaker Information Page available here.
If you need assistance please contact submissions@iscb.org and provide your poster title or submission ID.
Category K - 'Population Genetics Variation and Evolution'
Short Abstract: Egypt has become an epicenter of highly pathogenic avian influenza H5N1 influenza transmission. Like many viruses, the diffusion of H5N1 is a highly complicated process that depends on a large number of factors, most of which are poorly understood. We adopted a Bayesian phylogeographic GLM as developed by Lemey et al. in which viral diffusion patterns are reconstructed while predictors are simultaneously assessed. We analyzed a set of 226 sequences of the hemagglutinin gene from viruses identified as subclade 2.2.1.1., which was found almost exclusively within Egypt. We assessed each environmental and genetic predictor by calculating Bayes Factors (BFs) from the prior and posterior probabilities, with a significant cutoff score of 3. In addition to the GLM we analyzed cross-species transmission (CST) rates between each host species to estimate interactions from the widespread dispersal of H5N1 using Migrate-n. We found that predictors with high BF were exclusively from the governorate of origin including human density, the density of each avian host species, elevation, precipitation, latitude, and the genetic motif. CST values compliment the results of the GLM and indicate geese, ducks, and turkeys are primarily responsible for transmission to humans. This model could also be applied to different countries and viruses to inform response of public health agencies for dealing with pandemics.
Short Abstract: Geographically restricted, seasonal outbreaks of tularemia are widely prevalent in several regions of Sweden. The causative bacterium, Francisella tularensis, often spreads through different arthropod vectors causing ulceroglandular tularemia in humans. While the local ecological reservoirs of tularemia are yet to be ascertained, genetic and geographical analysis of the F. tularensis population could provide valuable insights into the point source of infection, migratory patterns, replication history, and population genetic parameters of the organism.
In this study, we analyzed the whole genome data of F. tularensis obtained from 134 individuals who contracted tularemia in several outbreaks between 1994 and 2010 in the Örebro County of Sweden. The genomic data was further compared with 29 reference genomes obtained from other regions of Sweden to study the phylogeny of the pathogen. The genomes were classified into different phylogenetic clades based on SNV markers.
The genomic diversity among the strains confirmed that several distinct subpopulations were simultaneously present within localized outbreaks. The data revealed little accumulation of mutation indicating very low rate of replication. Persistence of clones, within different subpopulations, with no spatial or temporal correlation was observed. We found one such clonal population to have disappeared in recent outbreaks consistent with the progressive purging of deleterious mutations by natural selection in bacterial populations.
Short Abstract: With the advance of high-throughput sequencing technologies, whole exome sequencing has been increasingly applied to identify disease-causing variants, especially the rare non-synonymous variants. Many statistical association tests and bioinformatics prediction tools have been proposed recently to identify those disease-causing genes/variants. Here we describe integrated methods to combine functional impact with statistical association. Our result shows that integrated methods perform consistently well under different simulation settings. They outperform statistical association tests when sample size and the number of disease-causing variants are moderate. Therefore, the integration of multiple sources of evidence for rare non-synonymous variants with respect to phenotype can identify disease gene and further help understand the disease mechanism.
Short Abstract: RNA viruses evolve quickly in response to selective pressure, emerging in novel hosts and causing outbreaks. We describe computational approaches for evaluating the effects of key factors influencing the relative ease with which a given virus population adapts to a novel host environment under serial passages. We modeled the evolutionary dynamics of viruses based on realistic descriptions of genomic changes as well as virus-host cell dynamics, and found that the likelihood of adaptation to new host cells depends most strongly on the genomic distance to more fit genotypes and the size of bottleneck selection. Using fitness functions connecting genotypes to phenotypes empirically inferred from sequences, we applied our approach to the adaptation of avian influenza A H5N1 viruses in mammalian hosts, which revealed that viral adaptation in reality is governed not only by fitness changes but also by randomness and accessibility of genotypes via mutations.
Short Abstract: Computational methods for detection of adaptation or positive selection in proteins are fundamental in our efforts to understand the proteomic basis of phenotypic evolution. However, predictions made by standard codon-based techniques for detection of selection (i.e., Ka/Ks) are often inconsistent with experimentally-derived sites of functional adaptation. Here, we present a novel approach for the prediction of protein adaptation that integrates structural and phylogenetic information to pinpoint evolutionary changes that are likely to have conferred significant structural and therefore functional change.
Maximum-likelihood methods are used to reconstruct ancestral sequences in a phylogenetic tree given a protein sequence alignment. Amino acid substitutions from ancestral-to-derived states are then evaluated structurally using physical information (solvent accessibility, catalytic sites, binding pockets, distances) acquired from multiple template structures from the Protein Data Bank (PDB). Structural properties of all residue substitutions as well as site-specific rates are then combined to calculate a likelihood of functional change, providing an estimate of structural adaptation.
To test this method, we have compared its predictions to previously identified sites of positive selection from the Selectome database, and achieve a significant correlation (r ~ 0.35, p < 0.01) with site-specific Ka/Ks predictions. We have applied the approach to reconstruct the structural evolution of pathogenesis-related proteins in Arabidopsis and detect enzymatic site and binding site adaptations more specifically than Ka/Ks methods. This method shows promise in protein family and proteome-wide studies of structural adaptation.
Short Abstract: Cardiovascular diseases and other common diseases are determined by the complex interplay between genetic and environmental factors, diet being the most relevant. Moreover, Nutritional Genomics has demonstrated differences in the response to diet depending on the genotype, but more research on gene-diet interactions is needed for personalization of diets. Although omic data are obtained by high-throughput procedures, the assessment of dietary intake is still very time consuming. Here we present a computational platform that integrates the automation of food questionnaires in nutritional studies as well as the analysis of gene-diet interactions in determining intermediate and disease phenotypes for thousands of participants. We developed a Web service to assist in the automation of food frequency questionnaires and 24-hours dietary recalls, using several nutrient data sources. A nutrient and food ontology was developed and food-nutrient results were customizable and exportable (SPSS, etc.) for statistical analysis of gene-diet interactions. For this task, we also developed a tool that uses standard statistical packages to build and fit the gene-diet interaction models by means of syntax scripts in predicting one or more continuous or dichotomic phenotypes. Nutrigenetic models including control for covariates are automatically created for each polymorphism. From the outputs, this tool extracts a selected set of statistical of parameters and places them in a spreadsheet. The tool allows editing the set of filter parameters (by p-values, etc.), as well as plotting selected results. We implemented our tool in real data obtained in two cardiovascular studies, and demonstrated the excellent performance of this platform.
Short Abstract: Insertion sequences (ISs) are transposable elements found widely in prokaryotes and they play an important evolutionary role by promoting gene acquisition among microbes. One important function is their influence on bacterial pathogenicity and virulence, as they provide means to facilitate plasmid integration and transfer of antibiotic resistance genes. For this reason, there have been many studies that characterize different types of ISs and discover underlying transposition mechanisms. However, rates of insertion, deletion and recombination of ISs have been uncharacterized as it is challenging to obtain the direct estimates. We extended our A-Bruijn graph based structural variant (SV) detection framework and applied it to whole genome shotgun sequencing data of more than 1,000 mutation accumulation lines of multiple derivative strains of Escherichia coli, and recovered insertion, deletion and recombination events involving ISs to obtain the rates. Our method works by constructing A-Bruijn graphs of the reference genome, where all instances of highly similar repeats are collapsed into an edge. Upon mapping paired-end reads onto the graphs, clustering discordant pairs with the use of depth-of-coverage information is performed in order to detect SVs involving repeats. Insertion was the most frequent event (1,180 events) but deletion (excision) was not found at all, other than the lost ISs due to large segment deletions mediated by recombination of two homologous ISs (116 events). Based on the recovery, we estimate the overall insertion rate to be ~4.00x10-4 insertions per genome per generation and the rate of recombination is 1/10th that of insertion.
Short Abstract: Recent advances in population genomics have revealed a large number of genomic variations in individuals, including those causing amino acid changes in protein. These data provide the statistics of amino acid changes with direction information from the reference allele to the alternative one and play an important role in understanding the evolution of proteins in the time span of human population. Here we examined more than 0.6 million amino acid variations in the NHLBI Exome Sequencing Project. Most of the amino acid changes occur between residues whose codons differ by one base by the genetic code, and amino acid changes caused by the change of CpG dinucleotide to TpG or CpA occur frequently. CpG-related amino acid changes are more abundant in the exomes than in cancer causing mutations, implying that codons containing CpG may tend to appear in the protein regions which are not critical in protein stability and function. Rare (<1%) variations included less CpG-related changes than common (>5%) variations, which may be due to larger functional influence of rare variations. To test whether codons with CpG are decreased in conserved residues, we examined the codon usages of arginine residues in strong conservation, such as those in the voltage sensors of the voltage-dependent ion channels. Unexpectedly, among the six codons of arginine, CGC was significantly abundant but AGA and AGA codons are decreased in voltage sensors. These results indicate the large impact of CpG mutability and its conservation in protein evolution.
View Posters By Category
- A) Bioinformatics of Disease and Treatment
- B) Comparative Genomics
- C) Education
- D) Epigenetics
- E) Functional Genomics
- F) Genome Organization and Annotation
- G) Genetic Variation Analysis
- H) Metagenomics
- I) Open Science and Citizen Science
- J) Pathogen informatics
- K) Population Genetics Variation and Evolution
- L) Protein Structure and Function Prediction and Analysis
- M) Proteomics
- N) Sequence Analysis
- O) Systems Biology and Networks
- P) Other
| Search Posters: |