Title: A comprehensive evaluation of long read error correction methods
Abstract Background Third-generation single molecule sequencing technologies can sequence long reads, which is advancing the frontiers of genomics research. However, their high error rates prohibit accurate and efficient downstream analysis. This difficulty has motivated the development of many long read error correction tools, which tackle this problem through sampling redundancy and/or leveraging accurate short reads of the same biological samples. Existing studies to asses these tools use simulated data sets, and are not sufficiently comprehensive in the range of software covered or diversity of evaluation measures used. Results In this paper, we present a categorization and review of long read error correction methods, and provide a comprehensive evaluation of the corresponding long read error correction tools. Leveraging recent real sequencing data, we establish benchmark data sets and set up evaluation criteria for a comparative assessment which includes quality of error correction as well as run-time and memory usage. We study how trimming and long read sequencing depth affect error correction in terms of length distribution and genome coverage post-correction, and the impact of error correction performance on an important application of long reads, genome assembly. We provide guidelines for practitioners for choosing among the available error correction tools and identify directions for future research. Conclusions Despite the high error rate of long reads, the state-of-the-art correction tools can achieve high correction quality. When short reads are available, the best hybrid methods outperform non-hybrid methods in terms of correction quality and computing resource usage. When choosing tools for use, practitioners are suggested to be careful with a few correction tools that discard reads, and check the effect of error correction tools on downstream analysis. Our evaluation code is available as open-source at https://github.com/haowenz/LRECE . more »« less
He, Cheng; Lin, Guifang; Wei, Hairong; Tang, Haibao; White, Frank F; Valent, Barbara; Liu, Sanzhen(
, NAR Genomics and Bioinformatics)
null
(Ed.)
Abstract Genome sequences provide genomic maps with a single-base resolution for exploring genetic contents. Sequencing technologies, particularly long reads, have revolutionized genome assemblies for producing highly continuous genome sequences. However, current long-read sequencing technologies generate inaccurate reads that contain many errors. Some errors are retained in assembled sequences, which are typically not completely corrected by using either long reads or more accurate short reads. The issue commonly exists, but few tools are dedicated for computing error rates or determining error locations. In this study, we developed a novel approach, referred to as k-mer abundance difference (KAD), to compare the inferred copy number of each k-mer indicated by short reads and the observed copy number in the assembly. Simple KAD metrics enable to classify k-mers into categories that reflect the quality of the assembly. Specifically, the KAD method can be used to identify base errors and estimate the overall error rate. In addition, sequence insertion and deletion as well as sequence redundancy can also be detected. Collectively, KAD is valuable for quality evaluation of genome assemblies and, potentially, provides a diagnostic tool to aid in precise error correction. KAD software has been developed to facilitate public uses.
The emergence of third‐generation sequencing (3GS; long‐reads) is bringing closer the goal of chromosome‐size fragments in de novo genome assemblies. This allows the exploration of new and broader questions on genome evolution for a number of nonmodel organisms. However, long‐read technologies result in higher sequencing error rates and therefore impose an elevated cost of sufficient coverage to achieve high enough quality. In this context, hybrid assemblies, combining short‐reads and long‐reads, provide an alternative efficient and cost‐effective approach to generate de novo, chromosome‐level genome assemblies. The array of available software programs for hybrid genome assembly, sequence correction and manipulation are constantly being expanded and improved. This makes it difficult for nonexperts to find efficient, fast and tractable computational solutions for genome assembly, especially in the case of nonmodel organisms lacking a reference genome or one from a closely related species. In this study, we review and test the most recent pipelines for hybrid assemblies, comparing the model organismDrosophila melanogasterto a nonmodel cactophilicDrosophila,D. mojavensis. We show that it is possible to achieve excellent contiguity on this nonmodel organism using thedbg2olcpipeline.
The emergence of third-generation single-molecule sequencing (TGS) technology has revolutionized the generation of long reads, which are essential for genome assembly and have been widely employed in sequencing the SARS-CoV-2 virus during the COVID-19 pandemic. Although long-read sequencing has been crucial in understanding the evolution and transmission of the virus, the high error rate associated with these reads can lead to inadequate genome assembly and downstream biological interpretation. In this study, we evaluate the accuracy and robustness of machine learning (ML) models using six different embedding techniques on SARS-CoV-2 error-incorporated genome sequences. Our analysis includes two types of error-incorporated genome sequences: those generated using simulation tools to emulate error profiles of long-read sequencing platforms and those generated by introducing random errors. We show that the spaced k-mers embedding method achieves high accuracy in classifying error-free SARS-CoV-2 genome sequences, and the spaced k-mers and weighted k-mers embedding methods are highly accurate in predicting error-incorporated sequences. The fixed-length vectors generated by these methods contribute to the high accuracy achieved. Our study provides valuable insights for researchers to effectively evaluate ML models and gain a better understanding of the approach for accurate identification of critical SARS-CoV-2 genome sequences.
Abstract For any genome-based research, a robust genome assembly is required. De novo assembly strategies have evolved with changes in DNA sequencing technologies and have been through at least three phases: i) short-read only, ii) short- and long-read hybrid, and iii) long-read only assemblies. Each of the phases has their own error model. We hypothesized that hidden scaffolding errors in short-read assembly and erroneous long-read contigs degrades the quality of short- and long-read hybrid assemblies. We assembled the genome of T. borchgrevinki from data generated during each of the three phases and assessed the quality problems we encountered. We developed strategies such as k-mer-assembled region replacement, parameter optimization, and long-read sampling to address the error models. We demonstrated that a k-mer based strategy improved short-read assemblies as measured by BUSCO while mate-pair libraries introduced hidden scaffolding errors and perturbed BUSCO scores. Further, we found that although hybrid assemblies can generate higher contiguity they tend to suffer from lower quality. In addition, we found long-read only assemblies can be optimized for contiguity by sub-sampling length-restricted raw reads. Our results indicate that long-read contig assembly is the current best choice and that assemblies from phase I and phase II were of lower quality.
INTRODUCTION One of the central applications of the human reference genome has been to serve as a baseline for comparison in nearly all human genomic studies. Unfortunately, many difficult regions of the reference genome have remained unresolved for decades and are affected by collapsed duplications, missing sequences, and other issues. Relative to the current human reference genome, GRCh38, the Telomere-to-Telomere CHM13 (T2T-CHM13) genome closes all remaining gaps, adds nearly 200 million base pairs (Mbp) of sequence, corrects thousands of structural errors, and unlocks the most complex regions of the human genome for scientific inquiry. RATIONALE We demonstrate how the T2T-CHM13 reference genome universally improves read mapping and variant identification in a globally diverse cohort. This cohort includes all 3202 samples from the expanded 1000 Genomes Project (1KGP), sequenced with short reads, as well as 17 globally diverse samples sequenced with long reads. By applying state-of-the-art methods for calling single-nucleotide variants (SNVs) and structural variants (SVs), we document the strengths and limitations of T2T-CHM13 relative to its predecessors and highlight its promise for revealing new biological insights within technically challenging regions of the genome. RESULTS Across the 1KGP samples, we found more than 1 million additional high-quality variants genome-wide using T2T-CHM13 than with GRCh38. Within previously unresolved regions of the genome, we identified hundreds of thousands of variants per sample—a promising opportunity for evolutionary and biomedical discovery. T2T-CHM13 improves the Mendelian concordance rate among trios and eliminates tens of thousands of spurious SNVs per sample, including a reduction of false positives in 269 challenging, medically relevant genes by up to a factor of 12. These corrections are in large part due to improvements to 70 protein-coding genes in >9 Mbp of inaccurate sequence caused by falsely collapsed or duplicated regions in GRCh38. Using the T2T-CHM13 genome also yields a more comprehensive view of SVs genome-wide, with a greatly improved balance of insertions and deletions. Finally, by providing numerous resources for T2T-CHM13 (including 1KGP genotypes, accessibility masks, and prominent annotation databases), our work will facilitate the transition to T2T-CHM13 from the current reference genome. CONCLUSION The vast improvements in variant discovery across samples of diverse ancestries position T2T-CHM13 to succeed as the next prevailing reference for human genetics. T2T-CHM13 thus offers a model for the construction and study of high-quality reference genomes from globally diverse individuals, such as is now being pursued through collaboration with the Human Pangenome Reference Consortium. As a foundation, our work underscores the benefits of an accurate and complete reference genome for revealing diversity across human populations. Genomic features and resources available for T2T-CHM13. Comparisons to GRCh38 reveal broad improvements in SNVs, indels, and SVs discovered across diverse human populations by means of short-read (1KGP) and long-read sequencing (LRS). These improvements are due to resolution of complex genomic loci (nonsyntenic and previously unresolved), duplication errors, and discordant haplotypes, including those in medically relevant genes.
Zhang, Haowen, Jain, Chirag, and Aluru, Srinivas. A comprehensive evaluation of long read error correction methods. Retrieved from https://par.nsf.gov/biblio/10286673. BMC Genomics 21.S6 Web. doi:10.1186/s12864-020-07227-0.
@article{osti_10286673,
place = {Country unknown/Code not available},
title = {A comprehensive evaluation of long read error correction methods},
url = {https://par.nsf.gov/biblio/10286673},
DOI = {10.1186/s12864-020-07227-0},
abstractNote = {Abstract Background Third-generation single molecule sequencing technologies can sequence long reads, which is advancing the frontiers of genomics research. However, their high error rates prohibit accurate and efficient downstream analysis. This difficulty has motivated the development of many long read error correction tools, which tackle this problem through sampling redundancy and/or leveraging accurate short reads of the same biological samples. Existing studies to asses these tools use simulated data sets, and are not sufficiently comprehensive in the range of software covered or diversity of evaluation measures used. Results In this paper, we present a categorization and review of long read error correction methods, and provide a comprehensive evaluation of the corresponding long read error correction tools. Leveraging recent real sequencing data, we establish benchmark data sets and set up evaluation criteria for a comparative assessment which includes quality of error correction as well as run-time and memory usage. We study how trimming and long read sequencing depth affect error correction in terms of length distribution and genome coverage post-correction, and the impact of error correction performance on an important application of long reads, genome assembly. We provide guidelines for practitioners for choosing among the available error correction tools and identify directions for future research. Conclusions Despite the high error rate of long reads, the state-of-the-art correction tools can achieve high correction quality. When short reads are available, the best hybrid methods outperform non-hybrid methods in terms of correction quality and computing resource usage. When choosing tools for use, practitioners are suggested to be careful with a few correction tools that discard reads, and check the effect of error correction tools on downstream analysis. Our evaluation code is available as open-source at https://github.com/haowenz/LRECE .},
journal = {BMC Genomics},
volume = {21},
number = {S6},
author = {Zhang, Haowen and Jain, Chirag and Aluru, Srinivas},
editor = {null}
}
Warning: Leaving National Science Foundation Website
You are now leaving the National Science Foundation website to go to a non-government website.
Website:
NSF takes no responsibility for and exercises no control over the views expressed or the accuracy of
the information contained on this site. Also be aware that NSF's privacy policy does not apply to this site.
