Scott Hazelhurst

Scientific computing using virtual high-performance computing: a case study using the Amazon elastic computing cloud

High-performance computing systems are important in scientific computing. Clusters of computer systems --- which range greatly in size --- are a common architecture for high-performance computing. Small, dedicated clusters are affordable and cost-effective, but may not be powerful enough for real applications. Larger dedicated systems are expensive in absolute terms and may be inefficient because many individual groups may not be able to provide sustained workload for the cluster. Shared systems are cost-effective, but then availability and access become a problem. An alternative model is that of a virtual cluste, as exemplified by Amazons Elastic Computing Cloud (EC2). This provides customers with storage and CPU power on an ondemand basis, and allows a researcher to dynamically build their own, dedicated cluster of computers when they need it. Used by commercial web services deployers, this technology can be used in scientific computing applications. This paper presents a case study of the use of EC2 for scientific computing. The case study concludes that EC2 provides a feasible, cost-effective model in many application areas.

Symbolic Trajectory Evaluation

data types can be declared in FL. This means that data can be represented and manipulated at a high level. This is important for making specifications understandable, and critical for overcoming the limitations of BDDs. FL provides built-in functions that are the interface to the STE engine. The use of a fully programmable script language is a key factor in implementing our verification methodology. It means that our tool can be simple, but through the use of a flexible interface, a user can verify a wide range of problems. On top of Voss’s facilities, we have implemented a simple theorem prover to implement the compositional theory presented in Section 5.3 – we have called this augmented system called VossProver. We have actually implemented a number of such tools, experimenting with style and functionality. The description presented here is a general description of one of the latest versions.

A hybrid verification approach: Getting deep into the design

One method of handling the computational complexity of the verification process is to combine the strengths of different approaches. We propose a hybrid verification technology combining symbolic trajectory evaluation with either symbolic model checking or SAT-based model checking. This reduces significantly the cost (both human and computing) of verifying circuits with complex initialisation, as well as simplifying proof development by enhancing verification productivity. The approach has been tested on current Intel designs.

H3ABioNet, a sustainable Pan-African Bioinformatics Network for Human Heredity and Health in Africa

The application of genomics technologies to medicine and biomedical research is increasing in popularity, made possible by new high-throughput genotyping and sequencing technologies and improved data analysis capabilities. Some of the greatest genetic diversity among humans, animals, plants, and microbiota occurs in Africa, yet genomic research outputs from the continent are limited. The Human Heredity and Health in Africa (H3Africa) initiative was established to drive the development of genomic research for human health in Africa, and through recognition of the critical role of bioinformatics in this process, spurred the establishment of H3ABioNet, a pan-African bioinformatics network for H3Africa. The limitations in bioinformatics capacity on the continent have been a major contributory factor to the lack of notable outputs in high-throughput biology research. Although pockets of high-quality bioinformatics teams have existed previously, the majority of research institutions lack experienced faculty who can train and supervise bioinformatics students. H3ABioNet aims to address this dire need, specifically in the area of human genetics and genomics, but knock-on effects are ensuring this extends to other areas of bioinformatics. Here, we describe the emergence of genomics research and the development of bioinformatics in Africa through H3ABioNet.

Genetic diversity in black South Africans from Soweto

BackgroundDue to the unparalleled genetic diversity of its peoples, Africa is attracting growing research attention. Several African populations have been assessed in global initiatives such as the International HapMap and 1000 Genomes Projects. Notably excluded, however, is the southern Africa region, which is inhabited predominantly by southeastern Bantu-speakers, currently suffering under the dual burden of infectious and non-communicable diseases. Limited reference data for these individuals hampers medical research and prevents thorough understanding of the underlying population substructure. Here, we present the most detailed exploration, to date, of genetic diversity in 94 unrelated southeastern Bantu-speaking South Africans, resident in urban Soweto (Johannesburg).ResultsParticipants were typed for ~4.3 million SNPs using the Illumina Omni5 beadchip. PCA and ADMIXTURE plots were used to compare the observed variation with that seen in selected populations worldwide. Results indicated that Sowetans, and other southeastern Bantu-speakers, are a clearly distinct group from other African populations previously investigated, reflecting a unique genetic history with small, but significant contributions from diverse sources. To assess the suitability of our sample as representative of Sowetans, we compared our results to participants in a larger rheumatoid arthritis case–control study. The control group showed good clustering with our sample, but among the cases were individuals who demonstrated notable admixture.ConclusionsSowetan population structure appears unique compared to other black Africans, and may have clinical implications. Our data represent a suitable reference set for southeastern Bantu-speakers, on par with a HapMap type reference population, and constitute a prelude to the Southern African Human Genome Programme.

A Reconfigurable Approach to Packet Filtering

Network packet classification is an important function for firewalls and filters. Packet classification based on transport-layer headers is widely used, and is specified by providing the filter with a list of rules. The cost of lookup may become a bottleneck in network performance. We present a novel technique for packet classification using FPGAs that exploits the reprogrammable nature of FPGAs. The rules are converted into a boolean expression which is directly implemented as a circuit on an FPGA. This approach is cheaper and simpler than previous hardware implementations, and we have had good experimental results.

Population-specific common SNPs reflect demographic histories and highlight regions of genomic plasticity with functional relevance

BackgroundPopulation differentiation is the result of demographic and evolutionary forces. Whole genome datasets from the 1000 Genomes Project (October 2012) provide an unbiased view of genetic variation across populations from Europe, Asia, Africa and the Americas. Common population-specific SNPs (MAF > 0.05) reflect a deep history and may have important consequences for health and wellbeing. Their interpretation is contextualised by currently available genome data.ResultsThe identification of common population-specific (CPS) variants (SNPs and SSV) is influenced by admixture and the sample size under investigation. Nine of the populations in the 1000 Genomes Project (2 African, 2 Asian (including a merged Chinese group) and 5 European) revealed that the African populations (LWK and YRI), followed by the Japanese (JPT) have the highest number of CPS SNPs, in concordance with their histories and given the populations studied. Using two methods, sliding 50-SNP and 5-kb windows, the CPS SNPs showed distinct clustering across large genome segments and little overlap of clusters between populations. iHS enrichment score and the population branch statistic (PBS) analyses suggest that selective sweeps are unlikely to account for the clustering and population specificity. Of interest is the association of clusters close to recombination hotspots. Functional analysis of genes associated with the CPS SNPs revealed over-representation of genes in pathways associated with neuronal development, including axonal guidance signalling and CREB signalling in neurones.ConclusionsCommon population-specific SNPs are non-randomly distributed throughout the genome and are significantly associated with recombination hotspots. Since the variant alleles of most CPS SNPs are the derived allele, they likely arose in the specific population after a split from a common ancestor. Their proximity to genes involved in specific pathways, including neuronal development, suggests evolutionary plasticity of selected genomic regions. Contrary to expectation, selective sweeps did not play a large role in the persistence of population-specific variation. This suggests a stochastic process towards population-specific variation which reflects demographic histories and may have some interesting implications for health and susceptibility to disease.

Compositional model checking of partially ordered state spaces

Symbolic trajectory evaluation (STE)--a model checking technique based on partial order representations of state spaces--has been shown to be an effective model checking technique for large circuit models. However, the temporal logic that it supports is restricted, and as with all verification techniques has significant performance limitations. The demand for verifying larger circuits, and the need for greater expressiveness requires that both these problems be examined. The thesis develops a suitable logical framework for model checking partially ordered state spaces: the temporal logic TL and its associated satisfaction relations, based on the quaternary logic

KABOOM! A new suffix array based algorithm for clustering expression data

{\cal Q}.

H3Africa AWI-Gen Collaborative Centre: a resource to study the interplay between genomic and environmental risk factors for cardiometabolic diseases in four sub-Saharan African countries

TL is appropriate for expressing the truth of propositions about partially ordered state spaces, and has suitable technical properties that allow STE to support a richer temporal logic. Using this framework, verification conditions called assertions are defined, a generalised version of STE is developed, and three STE-based algorithms are proposed for investigation. Advantages of this style of proof include: models of time are incorporated; circuits can be described at a low level; and correctness properties are expressed at a relatively high level. A primary contribution of the thesis is the development of a compositional theory for TL assertions. This compositional theory is supported by the partial order representation of state space. To show the practical use of the compositional theory, two prototype verification systems were constructed, integrating theorem proving and STE. Data is manipulated efficiently by using binary decision diagrams as well as symbolic data representation methods. Simple heuristics and a flexible interface reduce the human cost of verification. Experiments were undertaken using these prototypes, including verifying two circuits from the IFIP WG 10.5 Benchmark suite. These experiments showed that the generalised STE algorithms were effective, and that through the use of the compositional theory it is possible to verify very large circuits completely, including detailed timing properties.