Lessons learned: overcoming common challenges in reconstructing the SARS-CoV-2 genome from short-read sequencing data via CoVpipe2

Background

Accurate genome sequences form the basis for genomic surveillance programs, the added value of which was impressively demonstrated during the COVID-19 pandemic by tracing transmission chains, discovering new viral lineages and mutations, and assessing them for infectiousness and resistance to available treatments. Amplicon strategies employing Illumina sequencing have become widely established for variant detection and reference-based reconstruction of SARS-CoV-2 genomes, and are routine bioinformatics tasks. Yet, specific challenges arise when analyzing amplicon data, for example, when crucial and even lineage-determining mutations occur near primer sites.

Methods

We present CoVpipe2, a bioinformatics workflow developed at the Public Health Institute of Germany to reconstruct SARS-CoV-2 genomes based on short-read sequencing data accurately. The decisive factor here is the reliable, accurate, and rapid reconstruction of genomes, considering the specifics of the used sequencing protocol. Besides fundamental tasks like quality control, mapping, variant calling, and consensus generation, we also implemented additional features to ease the detection of mixed samples and recombinants.

Results

We highlight common pitfalls in primer clipping, detecting heterozygote variants, and dealing with low-coverage regions and deletions. We introduce CoVpipe2 to address the above challenges and have compared and successfully validated the pipeline against selected publicly available benchmark datasets. CoVpipe2 features high usability, reproducibility, and a modular design that specifically addresses the characteristics of short-read amplicon protocols but can also be used for whole-genome short-read sequencing data.

Conclusions

CoVpipe2 has seen multiple improvement cycles and is continuously maintained alongside frequently updated primer schemes and new developments in the scientific community. Our pipeline is easy to set up and use and can serve as a blueprint for other pathogens in the future due to its flexibility and modularity, providing a long-term perspective for continuous support. CoVpipe2 is written in Nextflow and is freely accessible from \href{https://github.com/rki-mf1/CoVpipe2}{github.com/rki-mf1/CoVpipe2} under the GPL3 license.

Calling Structural Variants with Confidence from Short-Read Data in Wild Bird Populations

Comprehensive characterization of structural variation in natural populations has only become feasible in the last decade. To investigate the population genomic nature of structural variation, reproducible and high-confidence structural variation callsets are first required. We created a population-scale reference of the genome-wide landscape of structural variation across 33 Nordic house sparrows (Passer domesticus). To produce a consensus callset across all samples using short-read data, we compare heuristic-based quality filtering and visual curation (Samplot/PlotCritic and Samplot-ML) approaches. We demonstrate that curation of structural variants is important for reducing putative false positives and that the time invested in this step outweighs the potential costs of analyzing short-read-discovered structural variation data sets that include many potential false positives. We find that even a lenient manual curation strategy (e.g. applied by a single curator) can reduce the proportion of putative false positives by up to 80%, thus enriching the proportion of high-confidence variants. Crucially, in applying a lenient manual curation strategy with a single curator, nearly all (>99%) variants rejected as putative false positives were also classified as such by a more stringent curation strategy using three additional curators. Furthermore, variants rejected by manual curation failed to reflect the expected population structure from SNPs, whereas variants passing curation did. Combining heuristic-based quality filtering with rapid manual curation of structural variants in short-read data can therefore become a time- and cost-effective first step for functional and population genomic studies requiring high-confidence structural variation callsets.

The Application of Long-Read Sequencing to Cancer

Cancer is a multifaceted disease arising from numerous genomic aberrations that have been identified as a result of advancements in sequencing technologies. While next-generation sequencing (NGS), which uses short reads, has transformed cancer research and diagnostics, it is limited by read length. Third-generation sequencing (TGS), led by the Pacific Biosciences and Oxford Nanopore Technologies platforms, employs long-read sequences, which have marked a paradigm shift in cancer research. Cancer genomes often harbour complex events, and TGS, with its ability to span large genomic regions, has facilitated their characterisation, providing a better understanding of how complex rearrangements affect cancer initiation and progression. TGS has also characterised the entire transcriptome of various cancers, revealing cancer-associated isoforms that could serve as biomarkers or therapeutic targets. Furthermore, TGS has advanced cancer research by improving genome assemblies, detecting complex variants, and providing a more complete picture of transcriptomes and epigenomes. This review focuses on TGS and its growing role in cancer research. We investigate its advantages and limitations, providing a rigorous scientific analysis of its use in detecting previously hidden aberrations missed by NGS. This promising technology holds immense potential for both research and clinical applications, with far-reaching implications for cancer diagnosis and treatment.

PlasmidEC and gplas2: an optimized short-read approach to predict and reconstruct antibiotic resistance plasmids in Escherichia coli

Accurate reconstruction of Escherichia coli antibiotic resistance gene (ARG) plasmids from Illumina sequencing data has proven to be a challenge with current bioinformatic tools. In this work, we present an improved method to reconstruct E. coli plasmids using short reads. We developed plasmidEC, an ensemble classifier that identifies plasmid-derived contigs by combining the output of three different binary classification tools. We showed that plasmidEC is especially suited to classify contigs derived from ARG plasmids with a high recall of 0.941. Additionally, we optimized gplas, a graph-based tool that bins plasmid-predicted contigs into distinct plasmid predictions. Gplas2 is more effective at recovering plasmids with large sequencing coverage variations and can be combined with the output of any binary classifier. The combination of plasmidEC with gplas2 showed a high completeness (median=0.818) and F1-Score (median=0.812) when reconstructing ARG plasmids and exceeded the binning capacity of the reference-based method MOB-suite. In the absence of long-read data, our method offers an excellent alternative to reconstruct ARG plasmids in E. coli.

Patchwork: Alignment-Based Retrieval and Concatenation of Phylogenetic Markers from Genomic Data

Low-coverage whole-genome sequencing (also known as "genome skimming") is becoming an increasingly affordable approach to large-scale phylogenetic analyses. While already routinely used to recover organellar genomes, genome skimming is rather rarely utilized for recovering single-copy nuclear markers. One reason might be that only few tools exist to work with this data type within a phylogenomic context, especially to deal with fragmented genome assemblies. We here present a new software tool called Patchwork for mining phylogenetic markers from highly fragmented short-read assemblies as well as directly from sequence reads. Patchwork is an alignment-based tool that utilizes the sequence aligner DIAMOND and is written in the programming language Julia. Homologous regions are obtained via a sequence similarity search, followed by a "hit stitching" phase, in which adjacent or overlapping regions are merged into a single unit. The novel sliding window algorithm trims away any noncoding regions from the resulting sequence. We demonstrate the utility of Patchwork by recovering near-universal single-copy orthologs within a benchmarking study, and we additionally assess the performance of Patchwork in comparison with other programs. We find that Patchwork allows for accurate retrieval of (putatively) single-copy genes from genome skimming data sets at different sequencing depths with high computational speed, outperforming existing software targeting similar tasks. Patchwork is released under the GNU General Public License version 3. Installation instructions, additional documentation, and the source code itself are all available via GitHub at https://github.com/fethalen/Patchwork.

Corrigendum: Evaluation of CircRNA sequence assembly methods using long reads

[This corrects the article DOI: 10.3389/fgene.2022.816825.].

ROCker Models for Reliable Detection and Typing of Short-Read Sequences Carrying β-Lactamase Genes

Identification of genes encoding β-lactamases (BLs) from short-read sequences remains challenging due to the high frequency of shared amino acid functional domains and motifs in proteins encoded by BL genes and related non-BL gene sequences. Divergent BL homologs can be frequently missed during similarity searches, which has important practical consequences for monitoring antibiotic resistance. To address this limitation, we built ROCker models that targeted broad classes (e.g., class A, B, C, and D) and individual families (e.g., TEM) of BLs and challenged them with mock 150-bp- and 250-bp-read data sets of known composition. ROCker identifies most-discriminant bit score thresholds in sliding windows along the sequence of the target protein sequence and hence can account for nondiscriminative domains shared by unrelated proteins. BL ROCker models showed a 0% false-positive rate (FPR), a 0% to 4% false-negative rate (FNR), and an up-to-50-fold-higher F1 score [2 × precision × recall/(precision + recall)] compared to alternative methods, such as similarity searches using BLASTx with various e-value thresholds and BL hidden Markov models, or tools like DeepARG, ShortBRED, and AMRFinder. The ROCker models and the underlying protein sequence reference data sets and phylogenetic trees for read placement are freely available through http://enve-omics.ce.gatech.edu/data/rocker-bla. Application of these BL ROCker models to metagenomics, metatranscriptomics, and high-throughput PCR gene amplicon data should facilitate the reliable detection and quantification of BL variants encoded by environmental or clinical isolates and microbiomes and more accurate assessment of the associated public health risk, compared to the current practice.

IMPORTANCE Resistance genes encoding β-lactamases (BLs) confer resistance to the widely prescribed antibiotic class β-lactams. Therefore, it is important to assess the prevalence of BL genes in clinical or environmental samples for monitoring the spreading of these genes into pathogens and estimating public health risk. However, detecting BLs in short-read sequence data is technically challenging. Our ROCker model-based bioinformatics approach showcases the reliable detection and typing of BLs in complex data sets and thus contributes toward solving an important problem in antibiotic resistance surveillance. The ROCker models developed substantially expand the toolbox for monitoring antibiotic resistance in clinical or environmental settings.

AMR-meta: a k-mer and metafeature approach to classify antimicrobial resistance from high-throughput short-read metagenomics data

BACKGROUND

Antimicrobial resistance (AMR) is a global health concern. High-throughput metagenomic sequencing of microbial samples enables profiling of AMR genes through comparison with curated AMR databases. However, the performance of current methods is often hampered by database incompleteness and the presence of homology/homoplasy with other non-AMR genes in sequenced samples.

RESULTS

We present AMR-meta, a database-free and alignment-free approach, based on k-mers, which combines algebraic matrix factorization into metafeatures with regularized regression. Metafeatures capture multi-level gene diversity across the main antibiotic classes. AMR-meta takes in reads from metagenomic shotgun sequencing and outputs predictions about whether those reads contribute to resistance against specific classes of antibiotics. In addition, AMR-meta uses an augmented training strategy that joins an AMR gene database with non-AMR genes (used as negative examples). We compare AMR-meta with AMRPlusPlus, DeepARG, and Meta-MARC, further testing their ensemble via a voting system. In cross-validation, AMR-meta has a median f-score of 0.7 (interquartile range, 0.2-0.9). On semi-synthetic metagenomic data-external test-on average AMR-meta yields a 1.3-fold hit rate increase over existing methods. In terms of run-time, AMR-meta is 3 times faster than DeepARG, 30 times faster than Meta-MARC, and as fast as AMRPlusPlus. Finally, we note that differences in AMR ontologies and observed variance of all tools in classification outputs call for further development on standardization of benchmarking data and protocols.

CONCLUSIONS

AMR-meta is a fast, accurate classifier that exploits non-AMR negative sets to improve sensitivity and specificity. The differences in AMR ontologies and the high variance of all tools in classification outputs call for the deployment of standard benchmarking data and protocols, to fairly compare AMR prediction tools.

MetaMLP: A Fast Word Embedding Based Classifier to Profile Target Gene Databases in Metagenomic Samples

The functional profile of metagenomic samples enables improved understanding of microbial populations in the environment. Such analysis consists of assigning short sequencing reads to a particular functional category. Normally, manually curated databases are used for functional assignment, and genes are arranged into different classes. Sequence alignment has been widely used to profile metagenomic samples against curated databases. However, this method is time consuming and requires high computational resources. While several alignment-free methods based on k-mer composition have been developed in recent years, they still require large amounts of computer main memory. In this article, MetaMLP (Metagenomics Machine Learning Profiler), a machine learning method that represents sequences as numerical vectors (embeddings) and uses a simple one hidden layer neural network to profile functional categories, is developed. Unlike other methods, MetaMLP enables partial matching by using a reduced alphabet to build sequence embeddings from full and partial k-mers. MetaMLP is able to identify a slightly larger number of reads compared with DIAMOND (one of the fastest sequence alignment methods), as well as to perform accurate predictions with 0.99 precision and 0.99 recall. MetaMLP can process 100M reads in ∼10 minutes on a laptop computer, which is 50 times faster than DIAMOND.

Complete Whole Genome Sequences of Escherichia coli Surrogate Strains and Comparison of Sequence Methods with Application to the Food Industry

In 2013, the U.S. Department of Agriculture Food Safety and Inspection Service (USDA-FSIS) began transitioning to whole genome sequencing (WGS) for foodborne disease outbreak- and recall-associated isolate identification of select bacterial species. While WGS offers greater precision, certain hurdles must be overcome before widespread application within the food industry is plausible. Challenges include diversity of sequencing platform outputs and lack of standardized bioinformatics workflows for data analyses. We sequenced DNA from USDA-FSIS approved, non-pathogenic E. coli surrogates and a derivative group of rifampicin-resistant mutants (rif^R) via both Oxford Nanopore MinION and Illumina MiSeq platforms to generate and annotate complete genomes. Genome sequences from each clone were assembled separately so long-read, short-read, and combined sequence assemblies could be directly compared. The combined sequence data approach provides more accurate completed genomes. The genomes from these isolates were verified to lack functional key E. coli elements commonly associated with pathogenesis. Genetic alterations known to confer rif^R were also identified. As the food industry adopts WGS within its food safety programs, these data provide completed genomes for commonly used surrogate strains, with a direct comparison of sequence platforms and assembly strategies relevant to research/testing workflows applicable for both processors and regulators.

Integrated hybrid de novo assembly technologies to obtain high-quality pig genome using short and long reads

With the rapid progress of sequencing technologies, various types of sequencing reads and assembly algorithms have been designed to construct genome assemblies. Although recent studies have attempted to evaluate the appropriate type of sequencing reads and algorithms for assembling high-quality genomes, it is still a challenge to set the correct combination for constructing animal genomes. Here, we present a comparative performance assessment of 14 assembly combinations-9 software programs with different short and long reads of Duroc pig. Based on the results of the optimization process for genome construction, we designed an integrated hybrid de novo assembly pipeline, HSCG, and constructed a draft genome for Duroc pig. Comparison between the new genome and Sus scrofa 11.1 revealed important breakpoints in two S. scrofa 11.1 genes. Our findings may provide new insights into the pan-genome analysis studies of agricultural animals, and the integrated assembly pipeline may serve as a guide for the assembly of other animal genomes.

SRPRISM (Single Read Paired Read Indel Substitution Minimizer): an efficient aligner for assemblies with explicit guarantees

BACKGROUND

Alignment of sequence reads generated by next-generation sequencing is an integral part of most pipelines analyzing next-generation sequencing data. A number of tools designed to quickly align a large volume of sequences are already available. However, most existing tools lack explicit guarantees about their output. They also do not support searching genome assemblies, such as the human genome assembly GRCh38, that include primary and alternate sequences and placement information for alternate sequences to primary sequences in the assembly.

FINDINGS

This paper describes SRPRISM (Single Read Paired Read Indel Substitution Minimizer), an alignment tool for aligning reads without splices. SRPRISM has features not available in most tools, such as (i) support for searching genome assemblies with alternate sequences, (ii) partial alignment of reads with a specified region of reads to be included in the alignment, (iii) choice of ranking schemes for alignments, and (iv) explicit criteria for search sensitivity. We compare the performance of SRPRISM to GEM, Kart, STAR, BWA-MEM, Bowtie2, Hobbes, and Yara using benchmark sets for paired and single reads of lengths 100 and 250 bp generated using DWGSIM. SRPRISM found the best results for most benchmark sets with error rate of up to ∼2.5% and GEM performed best for higher error rates. SRPRISM was also more sensitive than other tools even when sensitivity was reduced to improve run time performance.

CONCLUSIONS

We present SRPRISM as a flexible read mapping tool that provides explicit guarantees on results.

AUSPP: A universal short-read pre-processing package

There are many short-read aligners that can map short reads to a reference genome/sequence, and most of them can directly accept a FASTQ file as the input query file. However, the raw data usually need to be pre-processed. Few software programs specialize in pre-processing raw data generated by a variety of next-generation sequencing (NGS) technologies. Here, we present AUSPP, a Perl script-based pipeline for pre-processing and automatic mapping of NGS short reads. This pipeline encompasses quality control, adaptor trimming, collapsing of reads, structural RNA removal, length selection, read mapping, and normalized wiggle file creation. It facilitates the processing from raw data to genome mapping and is therefore a powerful tool for the steps before meta-analysis. Most importantly, since AUSPP has default processing pipeline settings for many types of NGS data, most of the time, users will simply need to provide the raw data and genome. AUSPP is portable and easy to install, and the source codes are freely available at https://github.com/highlei/AUSPP.

A critical comparison of technologies for a plant genome sequencing project

BACKGROUND

A high-quality genome sequence of any model organism is an essential starting point for genetic and other studies. Older clone-based methods are slow and expensive, whereas faster, cheaper short-read-only assemblies can be incomplete and highly fragmented, which minimizes their usefulness. The last few years have seen the introduction of many new technologies for genome assembly. These new technologies and associated new algorithms are typically benchmarked on microbial genomes or, if they scale appropriately, on larger (e.g., human) genomes. However, plant genomes can be much more repetitive and larger than the human genome, and plant biochemistry often makes obtaining high-quality DNA that is free from contaminants difficult. Reflecting their challenging nature, we observe that plant genome assembly statistics are typically poorer than for vertebrates.

RESULTS

Here, we compare Illumina short read, Pacific Biosciences long read, 10x Genomics linked reads, Dovetail Hi-C, and BioNano Genomics optical maps, singly and combined, in producing high-quality long-range genome assemblies of the potato species Solanum verrucosum. We benchmark the assemblies for completeness and accuracy, as well as DNA compute requirements and sequencing costs.

CONCLUSIONS

The field of genome sequencing and assembly is reaching maturity, and the differences we observe between assemblies are surprisingly small. We expect that our results will be helpful to other genome projects, and that these datasets will be used in benchmarking by assembly algorithm developers.

Statistical correction for functional metagenomic profiling of a microbial community with short NGS reads

By sequence homology search, the list of all the functions found and the counts of reads being aligned to them present the functional profile of a metagenomic sample. However, a significant obstacle has been observed in this approach due to the short read length associated with many next generation sequencing technologies. This includes artificial families, cross-annotations, length bias and conservation bias. The widely applied cutoff methods, such as BLAST E-value, are not able to solve the problems. Following the published successful procedures on the artificial families and the cross-annotation issue, we propose in this paper to use zero-truncated Poisson and Binomial (ZTP-Bin) hierarchical modelling to correct the length bias and the conservation bias. Goodness-of-fit of the modelling and cross-validation for the prediction using a bioinformatic simulated sample show the validity of this approach. Evaluated on an in vitro-simulated data set, the proposed modelling method outperforms other traditional methods. All three steps were then sequentially applied on real-life metagenomic samples to show that the proposed framework will lead to a more accurate functional profile of a short read metagenomic sample.

A Novel Method to Detect Bias in Short Read NGS Data

Detecting sources of bias in transcriptomic data is essential to determine signals of Biological significance. We outline a novel method to detect sequence specific bias in short read Next Generation Sequencing data. This is based on determining intra-exon correlations between specific motifs. This requires a mild assumption that short reads sampled from specific regions from the same exon will be correlated with each other. This has been implemented on Apache Spark and used to analyse two D. melanogaster eye-antennal disc data sets generated at the same laboratory. The wild type data set in drosophila indicates a variation due to motif GC content that is more significant than that found due to exon GC content. The software is available online and could be applied for cross-experiment transcriptome data analysis in eukaryotes.

Alignment of Short Reads: A Crucial Step for Application of Next-Generation Sequencing Data in Precision Medicine

Precision medicine or personalized medicine has been proposed as a modernized and promising medical strategy. Genetic variants of patients are the key information for implementation of precision medicine. Next-generation sequencing (NGS) is an emerging technology for deciphering genetic variants. Alignment of raw reads to a reference genome is one of the key steps in NGS data analysis. Many algorithms have been developed for alignment of short read sequences since 2008. Users have to make a decision on which alignment algorithm to use in their studies. Selection of the right alignment algorithm determines not only the alignment algorithm but also the set of suitable parameters to be used by the algorithm. Understanding these algorithms helps in selecting the appropriate alignment algorithm for different applications in precision medicine. Here, we review current available algorithms and their major strategies such as seed-and-extend and q-gram filter. We also discuss the challenges in current alignment algorithms, including alignment in multiple repeated regions, long reads alignment and alignment facilitated with known genetic variants.

riboFrame: An Improved Method for Microbial Taxonomy Profiling from Non-Targeted Metagenomics

Non-targeted metagenomics offers the unprecedented possibility of simultaneously investigate the microbial profile and the genetic capabilities of a sample by a direct analysis of its entire DNA content. The assessment of the microbial taxonomic composition is frequently obtained by mapping reads to genomic databases that, although growing, are still limited and biased. Here we present riboFrame, a novel procedure for microbial profiling based on the identification and classification of 16S rDNA sequences in non-targeted metagenomics datasets. Reads overlapping the 16S rDNA genes are identified using Hidden Markov Models and a taxonomic assignment is obtained by naïve Bayesian classification. All reads identified as ribosomal are coherently positioned in the 16S rDNA gene, allowing the use of the topology of the gene (i.e., the secondary structure and the location of variable regions) to guide the abundance analysis. We tested and verified the effectiveness of our method on simulated ribosomal data, on simulated metagenomes and on a real dataset. riboFrame exploits the taxonomic potentialities of the 16S rDNA gene in the context of non-targeted metagenomics, giving an accurate perspective on the microbial profile in metagenomic samples.

NucBase, an easy to use read mapper for small RNAs

BACKGROUND

High-throughput deep-sequencing technology has generated an unprecedented number of expressed sequence reads that offer the opportunity to get insight into biological systems. Several databases report the sequence of small regulatory RNAs which play a prominent role in the control of transposable elements (TE). However, the huge amount of data reported in these databases remains mostly unexplored because the available tools are hard for biologists to use.

RESULTS

Here we report NucBase, a new program designed to make an exhaustive search for sequence matches and to align short sequence reads from large nucleic acid databases to genomes or input sequences. NucBase includes a graphical interface which allows biologists to align sequences with ease and immediately visualize matched sequences, their number and their genomic position. NucBase identifies nucleic motives with strict identity to input sequences, and it capably finds candidates with one or several mismatches. It offers the opportunity to identify "core sequences" comprised of a chosen number of consecutive matching nucleotides. This software can be run locally on any Windows, Linux or Mac OS computer with 32-bit architecture compatibility.

CONCLUSIONS

Since this software is easy to use and can detect reads that were undetected by other software, we believe that it will be useful for biologists involved in the field of TE silencing by small non-coding RNAs. We hope NucBase will be useful for a larger community of researchers, since it makes exploration of small nucleic sequences in any organism much easier.

BM-map: Bayesian mapping of multireads for next-generation sequencing data

Next-generation sequencing (NGS) technology generates millions of short reads, which provide valuable information for various aspects of cellular activities and biological functions. A key step in NGS applications (e.g., RNA-Seq) is to map short reads to correct genomic locations within the source genome. While most reads are mapped to a unique location, a significant proportion of reads align to multiple genomic locations with equal or similar numbers of mismatches; these are called multireads. The ambiguity in mapping the multireads may lead to bias in downstream analyses. Currently, most practitioners discard the multireads in their analysis, resulting in a loss of valuable information, especially for the genes with similar sequences. To refine the read mapping, we develop a Bayesian model that computes the posterior probability of mapping a multiread to each competing location. The probabilities are used for downstream analyses, such as the quantification of gene expression. We show through simulation studies and RNA-Seq analysis of real life data that the Bayesian method yields better mapping than the current leading methods. We provide a C++ program for downloading that is being packaged into a user-friendly software.