PERGA: A Paired-End Read Guided De Novo Assembler for Extending Contigs Using SVM and Look Ahead Approach

Since the read lengths of high throughput sequencing (HTS) technologies are short, de novo assembly which plays significant roles in many applications remains a great challenge. Most of the state-of-the-art approaches base on de Bruijn graph strategy and overlap-layout strategy. However, these approaches which depend on k-mers or read overlaps do not fully utilize information of paired-end and single-end reads when resolving branches. Since they treat all single-end reads with overlapped length larger than a fix threshold equally, they fail to use the more confident long overlapped reads for assembling and mix up with the relative short overlapped reads. Moreover, these approaches have not been special designed for handling tandem repeats (repeats occur adjacently in the genome) and they usually break down the contigs near the tandem repeats. We present PERGA (Paired-End Reads Guided Assembler), a novel sequence-reads-guided de novo assembly approach, which adopts greedy-like prediction strategy for assembling reads to contigs and scaffolds using paired-end reads and different read overlap size ranging from O _max to O _min to resolve the gaps and branches. By constructing a decision model using machine learning approach based on branch features, PERGA can determine the correct extension in 99.7% of cases. When the correct extension cannot be determined, PERGA will try to extend the contig by all feasible extensions and determine the correct extension by using look-ahead approach. Many difficult-resolved branches are due to tandem repeats which are close in the genome. PERGA detects such different copies of the repeats to resolve the branches to make the extension much longer and more accurate. We evaluated PERGA on both Illumina real and simulated datasets ranging from small bacterial genomes to large human chromosome, and it constructed longer and more accurate contigs and scaffolds than other state-of-the-art assemblers. PERGA can be freely downloaded at https://github.com/hitbio/PERGA.     

Performance comparison of second- and third-generation sequencers using a bacterial genome with two chromosomes

Background

The availability of diverse second- and third-generation sequencing technologies enables the rapid determination of the sequences of bacterial genomes. However, identifying the sequencing technology most suitable for producing a finished genome with multiple chromosomes remains a challenge. We evaluated the abilities of the following three second-generation sequencers: Roche 454 GS Junior (GS Jr), Life Technologies Ion PGM (Ion PGM), and Illumina MiSeq (MiSeq) and a third-generation sequencer, the Pacific Biosciences RS sequencer (PacBio), by sequencing and assembling the genome of Vibrio parahaemolyticus, which consists of a 5-Mb genome comprising two circular chromosomes.

Results

We sequenced the genome of V. parahaemolyticus with GS Jr, Ion PGM, MiSeq, and PacBio and performed de novo assembly with several genome assemblers. Although GS Jr generated the longest mean read length of 418 bp among the second-generation sequencers, the maximum contig length of the best assembly from GS Jr was 165 kbp, and the number of contigs was 309. Single runs of Ion PGM and MiSeq produced data of considerably greater sequencing coverage, 279× and 1,927×, respectively. The optimized result for Ion PGM contained 61 contigs assembled from reads of 77× coverage, and the longest contig was 895 kbp in size. Those for MiSeq were 34 contigs, 58× coverage, and 733 kbp, respectively. These results suggest that higher coverage depth is unnecessary for a better assembly result. We observed that multiple rRNA coding regions were fragmented in the assemblies from the second-generation sequencers, whereas PacBio generated two exceptionally long contigs of 3,288,561 and 1,875,537 bps, each of which was from a single chromosome, with 73× coverage and mean read length 3,119 bp, allowing us to determine the absolute positions of all rRNA operons.

Conclusions

PacBio outperformed the other sequencers in terms of the length of contigs and reconstructed the greatest portion of the genome, achieving a genome assembly of “finished grade” because of its long reads. It showed the potential to assemble more complex genomes with multiple chromosomes containing more repetitive sequences.

Electronic supplementary material

The online version of this article (doi:10.1186/1471-2164-15-699) contains supplementary material, which is available to authorized users.     

Microarray meta-analysis database (M2DB): a uniformly pre-processed, quality controlled, and manually curated human clinical microarray database

Background

With the rapid expansion of DNA sequencing databases, it is now feasible to identify relevant information from prior sequencing projects and completed genomes and apply it to de novo sequencing of new organisms. As an example, this paper demonstrates how such extra information can be used to improve de novo assemblies by augmenting the overlapping step. Finding all pairs of overlapping reads is a key task in many genome assemblers, and to this end, highly efficient algorithms have been developed to find alignments in large collections of sequences. It is well known that due to repeated sequences, many aligned pairs of reads nevertheless do not overlap. But no overlapping algorithm to date takes a rigorous approach to separating aligned but non-overlapping read pairs from true overlaps.

Results

We present an approach that extends the Minimus assembler by a data driven step to classify overlaps as true or false prior to contig construction. We trained several different classification models within the Weka framework using various statistics derived from overlaps of reads available from prior sequencing projects. These statistics included percent mismatch and k-mer frequencies within the overlaps as well as a comparative genomics score derived from mapping reads to multiple reference genomes. We show that in real whole-genome sequencing data from the E. coli and S. aureus genomes, by providing a curated set of overlaps to the contigging phase of the assembler, we nearly doubled the median contig length (N50) without sacrificing coverage of the genome or increasing the number of mis-assemblies.

Conclusions

Machine learning methods that use comparative and non-comparative features to classify overlaps as true or false can be used to improve the quality of a sequence assembly.     

Selected Research Articles from the 2019 International Workshop on Computational Network Biology: Modeling,Analysis, and Control (CNB-MAC)

Background

Haplotypes, the ordered lists of single nucleotide variations that distinguish chromosomal sequences from their homologous pairs, may reveal an individual’s susceptibility to hereditary and complex diseases and affect how our bodies respond to therapeutic drugs. Reconstructing haplotypes of an individual from short sequencing reads is an NP-hard problem that becomes even more challenging in the case of polyploids. While increasing lengths of sequencing reads and insert sizes helps improve accuracy of reconstruction, it also exacerbates computational complexity of the haplotype assembly task. This has motivated the pursuit of algorithmic frameworks capable of accurate yet efficient assembly of haplotypes from high-throughput sequencing data.

Results

We propose a novel graphical representation of sequencing reads and pose the haplotype assembly problem as an instance of community detection on a spatial random graph. To this end, we construct a graph where each read is a node with an unknown community label associating the read with the haplotype it samples. Haplotype reconstruction can then be thought of as a two-step procedure: first, one recovers the community labels on the nodes (i.e., the reads), and then uses the estimated labels to assemble the haplotypes. Based on this observation, we propose ComHapDet – a novel assembly algorithm for diploid and ployploid haplotypes which allows both bialleleic and multi-allelic variants.

Conclusions

Performance of the proposed algorithm is benchmarked on simulated as well as experimental data obtained by sequencing Chromosome 5 of tetraploid biallelic Solanum-Tuberosum (Potato). The results demonstrate the efficacy of the proposed method and that it compares favorably with the existing techniques.

     

A fault-tolerant method for HLA typing with PacBio data

Background

Human leukocyte antigen (HLA) genes are critical genes involved in important biomedical aspects, including organ transplantation, autoimmune diseases and infectious diseases. The gene family contains the most polymorphic genes in humans and the difference between two alleles is only a single base pair substitution in many cases. The next generation sequencing (NGS) technologies could be used for high throughput HLA typing but in silico methods are still needed to correctly assign the alleles of a sample. Computer scientists have developed such methods for various NGS platforms, such as Illumina, Roche 454 and Ion Torrent, based on the characteristics of the reads they generate. However, the method for PacBio reads was less addressed, probably owing to its high error rates. The PacBio system has the longest read length among available NGS platforms, and therefore is the only platform capable of having exon 2 and exon 3 of HLA genes on the same read to unequivocally solve the ambiguity problem caused by the “phasing” issue.

Results

We proposed a new method BayesTyping1 to assign HLA alleles for PacBio circular consensus sequencing reads using Bayes’ theorem. The method was applied to simulated data of the three loci HLA-A, HLA-B and HLA-DRB1. The experimental results showed its capability to tolerate the disturbance of sequencing errors and external noise reads.

Conclusions

The BayesTyping1 method could overcome the problems of HLA typing using PacBio reads, which mostly arise from sequencing errors of PacBio reads and the divergence of HLA genes, to some extent.

Electronic supplementary material

The online version of this article (doi:10.1186/1471-2105-15-296) contains supplementary material, which is available to authorized users.     

The effects of read length,quality and quantity on microsatellite discovery and primer development: from Illumina to PacBio

The advent of next‐generation sequencing (NGS) technologies has transformed the way microsatellites are isolated for ecological and evolutionary investigations. Recent attempts to employ NGS for microsatellite discovery have used the 454, Illumina, and Ion Torrent platforms, but other methods including single‐molecule real‐time DNA sequencing (Pacific Biosciences or PacBio) remain viable alternatives. We outline a workflow from sequence quality control to microsatellite marker validation in three plant species using PacBio circular consensus sequencing (CCS). We then evaluate the performance of PacBio CCS in comparison with other NGS platforms for microsatellite isolation, through simulations that focus on variations in read length, read quantity and sequencing error rate. Although quality control of CCS reads reduced microsatellite yield by around 50%, hundreds of microsatellite loci that are expected to have improved conversion efficiency to functional markers were retrieved for each species. The simulations quantitatively validate the advantages of long reads and emphasize the detrimental effects of sequencing errors on NGS‐enabled microsatellite development. In view of the continuing improvement in read length on NGS platforms, sequence quality and the corresponding strategies of quality control will become the primary factors to consider for effective microsatellite isolation. Among current options, PacBio CCS may be optimal for rapid, small‐scale microsatellite development due to its flexibility in scaling sequencing effort, while platforms such as Illumina MiSeq will provide cost‐efficient solutions for multispecies microsatellite projects.     

NGmerge: merging paired-end reads via novel empirically-derived models of sequencing errors

Background

Advances in Illumina DNA sequencing technology have produced longer paired-end reads that increasingly have sequence overlaps. These reads can be merged into a single read that spans the full length of the original DNA fragment, allowing for error correction and accurate determination of read coverage. Extant merging programs utilize simplistic or unverified models for the selection of bases and quality scores for the overlapping region of merged reads.

Results

We first examined the baseline quality score - error rate relationship using sequence reads derived from PhiX. In contrast to numerous published reports, we found that the quality scores produced by Illumina were not substantially inflated above the theoretical values, once the reference genome was corrected for unreported sequence variants. The PhiX reads were then used to create empirical models of sequencing errors in overlapping regions of paired-end reads, and these models were incorporated into a novel merging program, NGmerge. We demonstrate that NGmerge corrects errors and ambiguous bases better than other merging programs, and that it assigns quality scores for merged bases that accurately reflect the error rates. Our results also show that, contrary to published analyses, the sequencing errors of paired-end reads are not independent.

Conclusions

We provide a free and open-source program, NGmerge, that performs better than existing read merging programs. NGmerge is available on GitHub (https://github.com/harvardinformatics/NGmerge) under the MIT License; it is written in C and supported on Linux.

     

SAP-A Sequence Mapping and Analyzing Program for Long Sequence Reads Alignment and Accurate Variants Discovery

The third-generation of sequencing technologies produces sequence reads of 1000 bp or more that may contain high polymorphism information. However, most currently available sequence analysis tools are developed specifically for analyzing short sequence reads. While the traditional Smith-Waterman (SW) algorithm can be used to map long sequence reads, its naive implementation is computationally infeasible. We have developed a new Sequence mapping and Analyzing Program (SAP) that implements a modified version of SW to speed up the alignment process. In benchmarks with simulated and real exon sequencing data and a real E. coli genome sequence data generated by the third-generation sequencing technologies, SAP outperforms currently available tools for mapping short and long sequence reads in both speed and proportion of captured reads. In addition, it achieves high accuracy in detecting SNPs and InDels in the simulated data. SAP is available at https://github.com/davidsun/SAP.     

Unlocking Short Read Sequencing for Metagenomics

Background

Different high-throughput nucleic acid sequencing platforms are currently available but a trade-off currently exists between the cost and number of reads that can be generated versus the read length that can be achieved.

Methodology/Principal Findings

We describe an experimental and computational pipeline yielding millions of reads that can exceed 200 bp with quality scores approaching that of traditional Sanger sequencing. The method combines an automatable gel-less library construction step with paired-end sequencing on a short-read instrument. With appropriately sized library inserts, mate-pair sequences can overlap, and we describe the SHERA software package that joins them to form a longer composite read.

Conclusions/Significance

This strategy is broadly applicable to sequencing applications that benefit from low-cost high-throughput sequencing, but require longer read lengths. We demonstrate that our approach enables metagenomic analyses using the Illumina Genome Analyzer, with low error rates, and at a fraction of the cost of pyrosequencing.     

MetAMOS: a modular and open source metagenomic assembly and analysis pipeline

We describe MetAMOS, an open source and modular metagenomic assembly and analysis pipeline. MetAMOS represents an important step towards fully automated metagenomic analysis, starting with next-generation sequencing reads and producing genomic scaffolds, open-reading frames and taxonomic or functional annotations. MetAMOS can aid in reducing assembly errors, commonly encountered when assembling metagenomic samples, and improves taxonomic assignment accuracy while also reducing computational cost. MetAMOS can be downloaded from: https://github.com/treangen/MetAMOS.     

DecGPU: distributed error correction on massively parallel graphics processing units using CUDA and MPI

Background  

Next-generation sequencing technologies have led to the high-throughput production of sequence data (reads) at low cost. However, these reads are significantly shorter and more error-prone than conventional Sanger shotgun reads. This poses a challenge for the de novo assembly in terms of assembly quality and scalability for large-scale short read datasets.     

单分子测序技术及应用研究进展

从DNA双螺旋结构的发现开始,生命科学研究进入分子水平,在20世纪70年代出现的测序技术为破译遗传密码作出了巨大贡献.近几年出现的单分子测序技术,可以在单个分子水平读取核苷酸序列,也被称为第三代测序技术,主要代表有HeliScope、Nanopore和PacBio等.与传统的第一代和第二代测序技术相比,第三代测序能够产生更长的碱基读长,能直接对RNA进行测序,无需逆转录,测序速度极快,同时其中某些技术所涉及的设备可以小型化,可便携至野外现场测序.第三代测序技术在生命科学基础理论研究及生物医学临床实践中,具有广泛的应用.本文重点介绍了各种单分子测序技术的原理、优缺点,及其应用研究进展.     

Metagenomics: Read Length Matters

Obtaining an unbiased view of the phylogenetic composition and functional diversity within a microbial community is one central objective of metagenomic analysis. New technologies, such as 454 pyrosequencing, have dramatically reduced sequencing costs, to a level where metagenomic analysis may become a viable alternative to more-focused assessments of the phylogenetic (e.g., 16S rRNA genes) and functional diversity of microbial communities. To determine whether the short (~100 to 200 bp) sequence reads obtained from pyrosequencing are appropriate for the phylogenetic and functional characterization of microbial communities, the results of BLAST and COG analyses were compared for long (~750 bp) and randomly derived short reads from each of two microbial and one virioplankton metagenome libraries. Overall, BLASTX searches against the GenBank nr database found far fewer homologs within the short-sequence libraries. This was especially pronounced for a Chesapeake Bay virioplankton metagenome library. Increasing the short-read sampling depth or the length of derived short reads (up to 400 bp) did not completely resolve the discrepancy in BLASTX homolog detection. Only in cases where the long-read sequence had a close homolog (low BLAST E-score) did the derived short-read sequence also find a significant homolog. Thus, more-distant homologs of microbial and viral genes are not detected by short-read sequences. Among COG hits, derived short reads sampled at a depth of two short reads per long read missed up to 72% of the COG hits found using long reads. Noting the current limitation in computational approaches for the analysis of short sequences, the use of short-read-length libraries does not appear to be an appropriate tool for the metagenomic characterization of microbial communities.     

Detecting genomic deletions from high-throughput sequence data with unsupervised learning

Background

Structural variation (SV), which ranges from 50 bp to \(\sim\) 3 Mb in size, is an important type of genetic variations. Deletion is a type of SV in which a part of a chromosome or a sequence of DNA is lost during DNA replication. Three types of signals, including discordant read-pairs, reads depth and split reads, are commonly used for SV detection from high-throughput sequence data. Many tools have been developed for detecting SVs by using one or multiple of these signals.

Results

In this paper, we develop a new method called EigenDel for detecting the germline submicroscopic genomic deletions. EigenDel first takes advantage of discordant read-pairs and clipped reads to get initial deletion candidates, and then it clusters similar candidates by using unsupervised learning methods. After that, EigenDel uses a carefully designed approach for calling true deletions from each cluster. We conduct various experiments to evaluate the performance of EigenDel on low coverage sequence data.

Conclusions

Our results show that EigenDel outperforms other major methods in terms of improving capability of balancing accuracy and sensitivity as well as reducing bias. EigenDel can be downloaded from https://github.com/lxwgcool/EigenDel.

     

Mining statistically-solid k-mers for accurate NGS error correction

Background

NGS data contains many machine-induced errors. The most advanced methods for the error correction heavily depend on the selection of solid k-mers. A solid k-mer is a k-mer frequently occurring in NGS reads. The other k-mers are called weak k-mers. A solid k-mer does not likely contain errors, while a weak k-mer most likely contains errors. An intensively investigated problem is to find a good frequency cutoff f₀ to balance the numbers of solid and weak k-mers. Once the cutoff is determined, a more challenging but less-studied problem is to: (i) remove a small subset of solid k-mers that are likely to contain errors, and (ii) add a small subset of weak k-mers, that are likely to contain no errors, into the remaining set of solid k-mers. Identification of these two subsets of k-mers can improve the correction performance.

Results

We propose to use a Gamma distribution to model the frequencies of erroneous k-mers and a mixture of Gaussian distributions to model correct k-mers, and combine them to determine f₀. To identify the two special subsets of k-mers, we use the z-score of k-mers which measures the number of standard deviations a k-mer’s frequency is from the mean. Then these statistically-solid k-mers are used to construct a Bloom filter for error correction. Our method is markedly superior to the state-of-art methods, tested on both real and synthetic NGS data sets.

Conclusion

The z-score is adequate to distinguish solid k-mers from weak k-mers, particularly useful for pinpointing out solid k-mers having very low frequency. Applying z-score on k-mer can markedly improve the error correction accuracy.

     

Efficient and accurate whole genome assembly and methylome profiling of E. coli

Background

With the price of next generation sequencing steadily decreasing, bacterial genome assembly is now accessible to a wide range of researchers. It is therefore necessary to understand the best methods for generating a genome assembly, specifically, which combination of sequencing and bioinformatics strategies result in the most accurate assemblies. Here, we sequence three E. coli strains on the Illumina MiSeq, Life Technologies Ion Torrent PGM, and Pacific Biosciences RS. We then perform genome assemblies on all three datasets alone or in combination to determine the best methods for the assembly of bacterial genomes.

Results

Three E. coli strains – BL21(DE3), Bal225, and DH5α – were sequenced to a depth of 100× on the MiSeq and Ion Torrent machines and to at least 125× on the PacBio RS. Four assembly methods were examined and compared. The previously published BL21(DE3) genome [GenBank:{"type":"entrez-nucleotide","attrs":{"text":"AM946981.2","term_id":"313848522","term_text":"AM946981.2"}}AM946981.2], allowed us to evaluate the accuracy of each of the BL21(DE3) assemblies. BL21(DE3) PacBio-only assemblies resulted in a 90% reduction in contigs versus short read only assemblies, while N50 numbers increased by over 7-fold. Strikingly, the number of SNPs in PacBio-only assemblies were less than half that seen with short read assemblies (~20 SNPs vs. ~50 SNPs) and indels also saw dramatic reductions (~2 indel >5 bp in PacBio-only assemblies vs. ~12 for short-read only assemblies). Assemblies that used a mixture of PacBio and short read data generally fell in between these two extremes. Use of PacBio sequencing reads also allowed us to call covalent base modifications for the three strains. Each of the strains used here had a known covalent base modification genotype, which was confirmed by PacBio sequencing.

Conclusion

Using data generated solely from the Pacific Biosciences RS, we were able to generate the most complete and accurate de novo assemblies of E. coli strains. We found that the addition of other sequencing technology data offered no improvements over use of PacBio data alone. In addition, the sequencing data from the PacBio RS allowed for sensitive and specific calling of covalent base modifications.

Electronic supplementary material

The online version of this article (doi:10.1186/1471-2164-14-675) contains supplementary material, which is available to authorized users.