Illumina error correction near highly repetitive DNA regions improves de novo genome assembly

Background

Several standalone error correction tools have been proposed to correct sequencing errors in Illumina data in order to facilitate de novo genome assembly. However, in a recent survey, we showed that state-of-the-art assemblers often did not benefit from this pre-correction step. We found that many error correction tools introduce new errors in reads that overlap highly repetitive DNA regions such as low-complexity patterns or short homopolymers, ultimately leading to a more fragmented assembly.

Results

We propose BrownieCorrector, an error correction tool for Illumina sequencing data that focuses on the correction of only those reads that overlap short DNA patterns that are highly repetitive in the genome. BrownieCorrector extracts all reads that contain such a pattern and clusters them into different groups using a community detection algorithm that takes into account both the sequence similarity between overlapping reads and their respective paired-end reads. Each cluster holds reads that originate from the same genomic region and hence each cluster can be corrected individually, thus providing a consistent correction for all reads within that cluster.

Conclusions

BrownieCorrector is benchmarked using six real Illumina datasets for different eukaryotic genomes. The prior use of BrownieCorrector improves assembly results over the use of uncorrected reads in all cases. In comparison with other error correction tools, BrownieCorrector leads to the best assembly results in most cases even though less than 2% of the reads within a dataset are corrected. Additionally, we investigate the impact of error correction on hybrid assembly where the corrected Illumina reads are supplemented with PacBio data. Our results confirm that BrownieCorrector improves the quality of hybrid genome assembly as well. BrownieCorrector is written in standard C++11 and released under GPL license. BrownieCorrector relies on multithreading to take advantage of multi-core/multi-CPU systems. The source code is available at https://github.com/biointec/browniecorrector.

Electronic supplementary material

The online version of this article (10.1186/s12859-019-2906-2) contains supplementary material, which is available to authorized users.

Transcriptomic-Proteomic Correlation in the Predation-Evoked Venom of the Cone Snail, Conus imperialis

Individual variation in animal venom has been linked to geographical location, feeding habit, season, size, and gender. Uniquely, cone snails possess the remarkable ability to change venom composition in response to predatory or defensive stimuli. To date, correlations between the venom gland transcriptome and proteome within and between individual cone snails have not been reported. In this study, we use 454 pyrosequencing and mass spectrometry to decipher the transcriptomes and proteomes of the venom gland and corresponding predation-evoked venom of two specimens of Conus imperialis. Transcriptomic analyses revealed 17 conotoxin gene superfamilies common to both animals, including 5 novel superfamilies and two novel cysteine frameworks. While highly expressed transcripts were common to both specimens, variation of moderately and weakly expressed precursor sequences was surprisingly diverse, with one specimen expressing two unique gene superfamilies and consistently producing more paralogs within each conotoxin gene superfamily. Using a quantitative labelling method, conotoxin variability was compared quantitatively, with highly expressed peptides showing a strong correlation between transcription and translation, whereas peptides expressed at lower levels showed a poor correlation. These results suggest that major transcripts are subject to stabilizing selection, while minor transcripts are subject to diversifying selection.

Assembly of chloroplast genomes with long- and short-read data: a comparison of approaches using Eucalyptus pauciflora as a test case

BACKGROUND

Chloroplasts are organelles that conduct photosynthesis in plant and algal cells. The information chloroplast genome contained is widely used in agriculture and studies of evolution and ecology. Correctly assembling chloroplast genomes can be challenging because the chloroplast genome contains a pair of long inverted repeats (10-30 kb). Typically, it is simply assumed that the gross structure of the chloroplast genome matches the most commonly observed structure of two single-copy regions separated by a pair of inverted repeats. The advent of long-read sequencing technologies should remove the need to make this assumption by providing sufficient information to completely span the inverted repeat regions. Yet, long-reads tend to have higher error rates than short-reads, and relatively little is known about the best way to combine long- and short-reads to obtain the most accurate chloroplast genome assemblies. Using Eucalyptus pauciflora, the snow gum, as a test case, we evaluated the effect of multiple parameters, such as different coverage of long-(Oxford nanopore) and short-(Illumina) reads, different long-read lengths, different assembly pipelines, with a view to determining the most accurate and efficient approach to chloroplast genome assembly.

RESULTS

Hybrid assemblies combining at least 20x coverage of both long-reads and short-reads generated a single contig spanning the entire chloroplast genome with few or no detectable errors. Short-read-only assemblies generated three contigs (the long single copy, short single copy and inverted repeat regions) of the chloroplast genome. These contigs contained few single-base errors but tended to exclude several bases at the beginning or end of each contig. Long-read-only assemblies tended to create multiple contigs with a much higher single-base error rate. The chloroplast genome of Eucalyptus pauciflora is 159,942 bp, contains 131 genes of known function.

CONCLUSIONS

Our results suggest that very accurate assemblies of chloroplast genomes can be achieved using a combination of at least 20x coverage of long- and short-reads respectively, provided that the long-reads contain at least ~5x coverage of reads longer than the inverted repeat region. We show that further increases in coverage give little or no improvement in accuracy, and that hybrid assemblies are more accurate than long-read-only or short-read-only assemblies.

The present and future of de novo whole-genome assembly

As the advent of next-generation sequencing (NGS) technology, various de novo assembly algorithms based on the de Bruijn graph have been developed to construct chromosome-level sequences. However, numerous technical or computational challenges in de novo assembly still remain, although many bright ideas and heuristics have been suggested to tackle the challenges in both experimental and computational settings. In this review, we categorize de novo assemblers on the basis of the type of de Bruijn graphs (Hamiltonian and Eulerian) and discuss the challenges of de novo assembly for short NGS reads regarding computational complexity and assembly ambiguity. Then, we discuss how the limitations of the short reads can be overcome by using a single-molecule sequencing platform that generates long reads of up to several kilobases. In fact, the long read assembly has caused a paradigm shift in whole-genome assembly in terms of algorithms and supporting steps. We also summarize (i) hybrid assemblies using both short and long reads and (ii) overlap-based assemblies for long reads and discuss their challenges and future prospects. This review provides guidelines to determine the optimal approach for a given input data type, computational budget or genome.

An efficient error correction algorithm using FM-index

BACKGROUND

High-throughput sequencing offers higher throughput and lower cost for sequencing a genome. However, sequencing errors, including mismatches and indels, may be produced during sequencing. Because, errors may reduce the accuracy of subsequent de novo assembly, error correction is necessary prior to assembly. However, existing correction methods still face trade-offs among correction power, accuracy, and speed.

RESULTS

We develop a novel overlap-based error correction algorithm using FM-index (called FMOE). FMOE first identifies overlapping reads by aligning a query read simultaneously against multiple reads compressed by FM-index. Subsequently, sequencing errors are corrected by k-mer voting from overlapping reads only. The experimental results indicate that FMOE has highest correction power with comparable accuracy and speed. Our algorithm performs better in long-read than short-read datasets when compared with others. The assembly results indicated different algorithms has its own strength and weakness, whereas FMOE is good for long or good-quality reads.

CONCLUSIONS

FMOE is freely available at https://github.com/ythuang0522/FMOC .

Evaluation of the impact of Illumina error correction tools on de novo genome assembly

BACKGROUND

Recently, many standalone applications have been proposed to correct sequencing errors in Illumina data. The key idea is that downstream analysis tools such as de novo genome assemblers benefit from a reduced error rate in the input data. Surprisingly, a systematic validation of this assumption using state-of-the-art assembly methods is lacking, even for recently published methods.

RESULTS

For twelve recent Illumina error correction tools (EC tools) we evaluated both their ability to correct sequencing errors and their ability to improve de novo genome assembly in terms of contig size and accuracy.

CONCLUSIONS

We confirm that most EC tools reduce the number of errors in sequencing data without introducing many new errors. However, we found that many EC tools suffer from poor performance in certain sequence contexts such as regions with low coverage or regions that contain short repeated or low-complexity sequences. Reads overlapping such regions are often ill-corrected in an inconsistent manner, leading to breakpoints in the resulting assemblies that are not present in assemblies obtained from uncorrected data. Resolving this systematic flaw in future EC tools could greatly improve the applicability of such tools.

Comparison of error correction algorithms for Ion Torrent PGM data: application to hepatitis B virus

Ion Torrent Personal Genome Machine (PGM) technology is a mid-length read, low-cost and high-speed next-generation sequencing platform with a relatively high insertion and deletion (indel) error rate. A full systematic assessment of the effectiveness of various error correction algorithms in PGM viral datasets (e.g., hepatitis B virus (HBV)) has not been performed. We examined 19 quality-trimmed PGM datasets for the HBV reverse transcriptase (RT) region and found a total error rate of 0.48% ± 0.12%. Deletion errors were clearly present at the ends of homopolymer runs. Tests using both real and simulated data showed that the algorithms differed in their abilities to detect and correct errors and that the error rate and sequencing depth significantly affected the performance. Of the algorithms tested, Pollux showed a better overall performance but tended to over-correct 'genuine' substitution variants, whereas Fiona proved to be better at distinguishing these variants from sequencing errors. We found that the combined use of Pollux and Fiona gave the best results when error-correcting Ion Torrent PGM viral data.

RECKONER: read error corrector based on KMC

Summary

Presence of sequencing errors in data produced by next-generation sequencers affects quality of downstream analyzes. Accuracy of them can be improved by performing error correction of sequencing reads. We introduce a new correction algorithm capable of processing eukaryotic close to 500 Mbp-genome-size, high error-rated data using less than 4 GB of RAM in about 35 min on 16-core computer.

Availability and Implementation

Program is freely available at http://sun.aei.polsl.pl/REFRESH/reckoner .

Contact

sebastian.deorowicz@polsl.pl.

Supplementary information

Supplementary data are available at Bioinformatics online.

Third generation sequencing technologies applied to diagnostic microbiology: benefits and challenges in applications and data analysis

INTRODUCTION

The diagnosis of infectious diseases is among the most successful areas of application of new generation sequencing technologies. The field has seen the development of numerous experimental and analytical approaches for the detection and the fine description of pathogenic and non-pathogenic microorganisms.

AREAS COVERED

Without claiming to be exhaustive with respect to all applications and methods developed over the years, this review focuses on the advantages and the issues brought by the new technologies, with an eye in particular to third generation sequencing methods. Both experimental procedures and algorithmic strategies are presented, following the most relevant publications which have led to progress in our ability of detecting infectious agents. Expert commentary: The technical advance brought by third generation sequencing platforms has the potential to significantly expand the range of diagnostic tools that will be available to clinicians. Nonetheless, the implementation of these technologies in clinical practice is still far from being actionable and will temporally follow the path undertaken by second generation methods, which still require the setup of standardized pipelines in both wet and dry laboratory procedures.

Jabba: hybrid error correction for long sequencing reads

Background

Third generation sequencing platforms produce longer reads with higher error rates than second generation technologies. While the improved read length can provide useful information for downstream analysis, underlying algorithms are challenged by the high error rate. Error correction methods in which accurate short reads are used to correct noisy long reads appear to be attractive to generate high-quality long reads. Methods that align short reads to long reads do not optimally use the information contained in the second generation data, and suffer from large runtimes. Recently, a new hybrid error correcting method has been proposed, where the second generation data is first assembled into a de Bruijn graph, on which the long reads are then aligned.

Results

In this context we present Jabba, a hybrid method to correct long third generation reads by mapping them on a corrected de Bruijn graph that was constructed from second generation data. Unique to our method is the use of a pseudo alignment approach with a seed-and-extend methodology, using maximal exact matches (MEMs) as seeds. In addition to benchmark results, certain theoretical results concerning the possibilities and limitations of the use of MEMs in the context of third generation reads are presented.

Conclusion

Jabba produces highly reliable corrected reads: almost all corrected reads align to the reference, and these alignments have a very high identity. Many of the aligned reads are error-free. Additionally, Jabba corrects reads using a very low amount of CPU time. From this we conclude that pseudo alignment with MEMs is a fast and reliable method to map long highly erroneous sequences on a de Bruijn graph.

Iterative error correction of long sequencing reads maximizes accuracy and improves contig assembly

Next-generation sequencers such as Illumina can now produce reads up to 300 bp with high throughput, which is attractive for genome assembly. A first step in genome assembly is to computationally correct sequencing errors. However, correcting all errors in these longer reads is challenging. Here, we show that reads with remaining errors after correction often overlap repeats, where short erroneous k-mers occur in other copies of the repeat. We developed an iterative error correction pipeline that runs the previously published String Graph Assembler (SGA) in multiple rounds of k-mer-based correction with an increasing k-mer size, followed by a final round of overlap-based correction. By combining the advantages of small and large k-mers, this approach corrects more errors in repeats and minimizes the total amount of erroneous reads. We show that higher read accuracy increases contig lengths two to three times. We provide SGA-Iteratively Correcting Errors (https://github.com/hillerlab/IterativeErrorCorrection/) that implements iterative error correction by using modules from SGA.