BackgroundHigh-throughput sequencing technologies, such as the Illumina Genome Analyzer, are powerful new tools for investigating a wide range of biological and medical questions. Statistical and computational methods are key for drawing meaningful and accurate conclusions from the massive and complex datasets generated by the sequencers. We provide a detailed evaluation of statistical methods for normalization and differential expression (DE) analysis of Illumina transcriptome sequencing (mRNA-Seq) data.ResultsWe compare statistical methods for detecting genes that are significantly DE between two types of biological samples and find that there are substantial differences in how the test statistics handle low-count genes. We evaluate how DE results are affected by features of the sequencing platform, such as, varying gene lengths, base-calling calibration method (with and without phi X control lane), and flow-cell/library preparation effects. We investigate the impact of the read count normalization method on DE results and show that the standard approach of scaling by total lane counts (e.g., RPKM) can bias estimates of DE. We propose more general quantile-based normalization procedures and demonstrate an improvement in DE detection.ConclusionsOur results have significant practical and methodological implications for the design and analysis of mRNA-Seq experiments. They highlight the importance of appropriate statistical methods for normalization and DE inference, to account for features of the sequencing platform that could impact the accuracy of results. They also reveal the need for further research in the development of statistical and computational methods for mRNA-Seq.

The search to understand how genomes innovate in response to selection dominates the field of evolutionary biology. Powerful molecular evolution approaches have been developed to test individual loci for signatures of selection. In many cases, however, an organism's response to changes in selective pressure may be mediated by multiple genes, whose products function together in a cellular process or pathway. Here we assess the prevalence of polygenic evolution in pathways in the yeasts Saccharomyces cerevisiae and S. bayanus. We first established short-read sequencing methods to detect cis-regulatory variation in a diploid hybrid between the species. We then tested for the scenario in which selective pressure in one species to increase or decrease the activity of a pathway has driven the accumulation of cis-regulatory variants that act in the same direction on gene expression. Application of this test revealed a variety of yeast pathways with evidence for directional regulatory evolution. In parallel, we also used population genomic sequencing data to compare protein and cis-regulatory variation within and between species. We identified pathways with evidence for divergence within S. cerevisiae, and we detected signatures of positive selection between S. cerevisiae and S. bayanus. Our results point to polygenic, pathway-level change as a common evolutionary mechanism among yeasts. We suggest that pathway analyses, including our test for directional regulatory evolution, will prove to be a relevant and powerful strategy in many evolutionary genomic applications.A main challenge of evolutionary biology is to understand the influence of selection on genetic variation within and between species. Classically, molecular evolution methods have targeted individual genes or loci for tests of selection. Their successes have uncovered both protein-coding and regulatory variants with signatures of non-neutral evolution (1-3). However, decades of quantitative genetic mapping studies indicate that in most cases, variation in phenotype between individuals is the result of multiple sequence changes at unlinked loci (4). The genetic response to changes in selective pressure is likely to follow the same pattern, but, to date, methods for identifying cases of polygenic evolution have been at a premium.Some of the strongest evidence for polygenic adaptation has emerged from the study of protein-coding variants between phylogenetic lineages. Signatures of positive selection can be detected in the protein-coding sequences of groups of genes of related function (5-8), suggestive of a coherent series of genetic changes accumulated by a population in response to selection. Additionally, genetic variation in gene expression represents a rich data source for signatures of selection, both positive (9, 10) and purifying (11,12), although the mechanisms that govern regulatory evolution are not fully understood. Regulatory variants can act in cis, to impact the expression of a neighboring gene, or in trans, targeting the expression of genes el...

Background: Biological studies involve a growing number of distinct high-throughput experiments to characterize samples of interest. There is a lack of methods to visualize these different genomic datasets in a versatile manner. In addition, genomic data analysis requires integrated visualization of experimental data along with constantly changing genomic annotation and statistical analyses.

A complete description of the transcriptome of an organism is crucial for a comprehensive understanding of how it functions and how its transcriptional networks are controlled, and may provide insights into the organism's evolution. Despite the status of Saccharomyces cerevisiae as arguably the most well-studied model eukaryote, we still do not have a full catalog or understanding of all its genes. In order to interrogate the transcriptome of S. cerevisiae for low abundance or rapidly turned over transcripts, we deleted elements of the RNA degradation machinery with the goal of preferentially increasing the relative abundance of such transcripts. We then used high-resolution tiling microarrays and ultra high–throughput sequencing (UHTS) to identify, map, and validate unannotated transcripts that are more abundant in the RNA degradation mutants relative to wild-type cells. We identified 365 currently unannotated transcripts, the majority presumably representing low abundance or short-lived RNAs, of which 185 are previously unknown and unique to this study. It is likely that many of these are cryptic unstable transcripts (CUTs), which are rapidly degraded and whose function(s) within the cell are still unclear, while others may be novel functional transcripts. Of the 185 transcripts we identified as novel to our study, greater than 80 percent come from regions of the genome that have lower conservation scores amongst closely related yeast species than 85 percent of the verified ORFs in S. cerevisiae. Such regions of the genome have typically been less well-studied, and by definition transcripts from these regions will distinguish S. cerevisiae from these closely related species.

DNA sequencing of PCR-amplified marker genes, especially but not limited to the 16S rRNA gene, is perhaps the most common approach for profiling microbial communities. Due to technological constraints of commonly available DNA sequencing, these approaches usually take the form of short reads sequenced from a narrow, targeted variable region, with a corresponding loss of taxonomic resolution relative to the full length marker gene. We use Pacific Biosciences single-molecule, real-time circular consensus sequencing to sequence amplicons spanning the entire length of the 16S rRNA gene. However, this sequencing technology suffers from high sequencing error rate that needs to be addressed in order to take full advantage of the longer sequence. Here, we present a method to model the sequencing error process using a generalized pair hidden Markov chain model and estimate bacterial abundances in microbial samples. We demonstrate, with simulated and real data, that our model and its associated estimation procedure are able to give accurate estimates at the species (or subspecies) level, and is more flexible than existing methods like SImple Non-Bayesian TAXonomy (SINTAX).