BACKGROUND: It is possible to predict whether a tuberculosis (TB) patient will fail to respond to specific antibiotics by sequencing the genome of the infecting Mycobacterium tuberculosis (Mtb) and observing whether the pathogen carries specific mutations at drug-resistance sites. This advancement has led to the collation of TB databases such as PATRIC and ReSeqTB that possess both whole genome sequences and drug resistance phenotypes of infecting Mtb isolates. Bioinformatics tools have also been developed to predict drug resistance from whole genome sequencing (WGS) data. Here, we evaluate the performance of four popular tools (TBProfiler, MyKrobe, KvarQ, PhyResSE) with 6746 isolates compiled from publicly available databases, and subsequently identify highly probable phenotyping errors in the databases by genetically predicting the drug phenotypes using all four software. RESULTS: Our results show that these bioinformatics tools generally perform well in predicting the resistance status for two key first-line agents (isoniazid, rifampicin), but the accuracy is lower for second-line injectables and fluoroquinolones. The error rates in the databases are also non-trivial, reaching as high as 31.1% for prothionamide, and that phenotypes from ReSeqTB are more susceptible to errors. CONCLUSIONS: The good performance of the automated software for drug resistance prediction from TB WGS data shown in this study further substantiates the usefulness and promise of utilising genetic data to accurately profile TB drug resistance, thereby reducing misdiagnoses arising from error-prone culture-based drug susceptibility testing.

MOTIVATION: Whole-genome sequencing (WGS) is now routinely used for the detection and identification of genetic variants, particularly single nucleotide polymorphisms (SNPs) in humans, and this has provided valuable new insights into human diversity, population histories and genetic association studies of traits and diseases. However, this relies on accurate detection and genotyping calling of the polymorphisms present in the samples sequenced. To minimize cost, the majority of current WGS studies, including the 1000 Genomes Project (1 KGP) have adopted low coverage sequencing of large number of samples, where such designs have inadvertently influenced the development of variant calling methods on WGS data. Assessment of variant accuracy are usually performed on the same set of low coverage individuals or a smaller number of deeply sequenced individuals. It is thus unclear how these variant calling methods would fare for a dataset of ∼100 samples from a population not part of the 1 KGP that have been sequenced at various coverage depths. AVAILABILITY AND IMPLEMENTATION: Using down-sampling of the sequencing reads obtained from the Singapore Sequencing Malay Project (SSMP), and a set of SNP calls from the same individuals genotyped on the Illumina Omni1-Quad array, we assessed the sensitivity of SNP detection, accuracy of genotype calls made and variant accuracy for six commonly used variant calling methods of GATK, SAMtools, Consensus Assessment of Sequence and Variation (CASAVA), VarScan, glfTools and SOAPsnp. The results indicate that at 5× coverage depth, the multi-sample callers of GATK and SAMtools yield the best accuracy particularly if the study samples are called together with a large number of individuals such as those from 1000 Genomes Project. If study samples are sequenced at a high coverage depth such as 30×, CASAVA has the highest variant accuracy as compared with the other variant callers assessed.

MOTIVATION: Next-generation genotyping microarrays have been designed with insights from 1000 Genomes Project and whole-exome sequencing studies. These arrays additionally include variants that are typically present at lower frequencies. Determining the genotypes of these variants from hybridization intensities is challenging because there is less support to locate the presence of the minor alleles when the allele counts are low. Existing algorithms are mainly designed for calling common variants and are notorious for failing to generate accurate calls for low-frequency and rare variants. Here, we introduce a new calling algorithm, iCall, to call genotypes for variants across the whole spectrum of allele frequencies. RESULTS: We benchmarked iCall against four of the most commonly used algorithms, GenCall, optiCall, illuminus and GenoSNP, as well as a post-processing caller zCall that adopted a two-stage calling design. Normalized hybridization intensities for 12 370 individuals genotyped on the Illumina HumanExome BeadChip were considered, of which 81 individuals were also whole-genome sequenced. The sequence calls were used to benchmark the accuracy of the genotype calling, and our comparisons indicated that iCall outperforms all four single-stage calling algorithms in terms of call rates and concordance, particularly in the calling accuracy of minor alleles, which is the principal concern for rare and low-frequency variants. The application of zCall to post-process the output from iCall also produced marginally improved performance to the combination of zCall and GenCall. AVAILABILITY AND IMPLEMENTATION: iCall is implemented in C++ for use on Linux operating systems and is available for download at http://www.statgen.nus.edu.sg/∼software/icall.html.

BACKGROUND: Many potentially life-threatening infectious viruses are highly mutable in nature. Characterizing the fittest variants within a quasispecies from infected patients is expected to allow unprecedented opportunities to investigate the relationship between quasispecies diversity and disease epidemiology. The advent of next-generation sequencing technologies has allowed the study of virus diversity with high-throughput sequencing, although these methods come with higher rates of errors which can artificially increase diversity. RESULTS: Here we introduce a novel computational approach that incorporates base quality scores from next-generation sequencers for reconstructing viral genome sequences that simultaneously infers the number of variants within a quasispecies that are present. Comparisons on simulated and clinical data on dengue virus suggest that the novel approach provides a more accurate inference of the underlying number of variants within the quasispecies, which is vital for clinical efforts in mapping the within-host viral diversity. Sequence alignments generated by our approach are also found to exhibit lower rates of error. CONCLUSIONS: The ability to infer the viral quasispecies colony that is present within a human host provides the potential for a more accurate classification of the viral phenotype. Understanding the genomics of viruses will be relevant not just to studying how to control or even eradicate these viral infectious diseases, but also in learning about the innate protection in the human host against the viruses.

MOTIVATION: Methods for detecting positive selection relied on finding evidence of long haplotypes to identify candidate regions under selection. However, these methods generally do not identify the length and form of the selected haplotype. RESULTS: We present HapFinder, a method which can find the common longest haplotype under three different settings from a database, which is relevant in the analysis of positive selection in population genetics and also in medical genetics for finding the likely haplotype form carrying the causal allele at the functional polymorphism. AVAILABILITY: A java program, implementing the methods described in HapFinder, together with R scripts and datasets for producing the figures presented in this article are publicly available at http://www.nus-cme.org.sg/sgvp/software/hapfinder.html. The site also hosts an online browser for finding haplotypes from the International HapMap Project and the Singapore Genome Variation Project.

MOTIVATION: Linkage disequilibrium (LD) differences between populations can potentially result in failure to replicate primary signals of trait association in independent genome-wide association studies (GWAS). However, such inter-population LD differences can be leveraged to narrow the search for common causal variants responsible for the association signals observed. The ability to assess and quantify LD variations among populations is thus expected to contribute to both replication and fine-mapping stages of GWAS. AVAILABILITY: The program varLD is available for download from http://www.nus-cme.org.sg/software/varld.html.