skip to main content


Title: EvalDNA: a machine learning-based tool for the comprehensive evaluation of mammalian genome assembly quality
Abstract Background

To select the most complete, continuous, and accurate assembly for an organism of interest, comprehensive quality assessment of assemblies is necessary. We present a novel tool, called Evaluation of De Novo Assemblies (EvalDNA), which uses supervised machine learning for the quality scoring of genome assemblies and does not require an existing reference genome for accuracy assessment.

Results

EvalDNA calculates a list of quality metrics from an assembled sequence and applies a model created from supervised machine learning methods to integrate various metrics into a comprehensive quality score. A well-tested, accurate model for scoring mammalian genome sequences is provided as part of EvalDNA. This random forest regression model evaluates an assembled sequence based on continuity, completeness, and accuracy, and was able to explain 86% of the variation in reference-based quality scores within the testing data. EvalDNA was applied to human chromosome 14 assemblies from the GAGE study to rank genome assemblers and to compare EvalDNA to two other quality evaluation tools. In addition, EvalDNA was used to evaluate several genome assemblies of the Chinese hamster genome to help establish a better reference genome for the biopharmaceutical manufacturing community. EvalDNA was also used to assess more recent human assemblies from the QUAST-LG study completed in 2018, and its ability to score bacterial genomes was examined through application on bacterial assemblies from the GAGE-B study.

Conclusions

EvalDNA enables scientists to easily identify the best available genome assembly for their organism of interest without requiring a reference assembly. EvalDNA sets itself apart from other quality assessment tools by producing a quality score that enables direct comparison among assemblies from different species.

 
more » « less
Award ID(s):
1736123
NSF-PAR ID:
10360389
Author(s) / Creator(s):
;
Publisher / Repository:
Springer Science + Business Media
Date Published:
Journal Name:
BMC Bioinformatics
Volume:
22
Issue:
1
ISSN:
1471-2105
Format(s):
Medium: X
Sponsoring Org:
National Science Foundation
More Like this
  1. Abstract Background

    Ambitious initiatives to coordinate genome sequencing of Earth's biodiversity mean that the accumulation of genomic data is growing rapidly. In addition to cataloguing biodiversity, these data provide the basis for understanding biological function and evolution. Accurate and complete genome assemblies offer a comprehensive and reliable foundation upon which to advance our understanding of organismal biology at genetic, species, and ecosystem levels. However, ever-changing sequencing technologies and analysis methods mean that available data are often heterogeneous in quality. To guide forthcoming genome generation efforts and promote efficient prioritization of resources, it is thus essential to define and monitor taxonomic coverage and quality of the data.

    Findings

    Here we present an automated analysis workflow that surveys genome assemblies from the United States NCBI, assesses their completeness using the relevant BUSCO datasets, and collates the results into an interactively browsable resource. We apply our workflow to produce a community resource of available assemblies from the phylum Arthropoda, the Arthropoda Assembly Assessment Catalogue. Using this resource, we survey current taxonomic coverage and assembly quality at the NCBI, examine how key assembly metrics relate to gene content completeness, and compare results from using different BUSCO lineage datasets.

    Conclusions

    These results demonstrate how the workflow can be used to build a community resource that enables large-scale assessments to survey species coverage and data quality of available genome assemblies, and to guide prioritizations for ongoing and future sampling, sequencing, and genome generation initiatives.

     
    more » « less
  2. Abstract Background

    Systems-level analyses, such as differential gene expression analysis, co-expression analysis, and metabolic pathway reconstruction, depend on the accuracy of the transcriptome. Multiple tools exist to perform transcriptome assembly from RNAseq data. However, assembling high quality transcriptomes is still not a trivial problem. This is especially the case for non-model organisms where adequate reference genomes are often not available. Different methods produce different transcriptome models and there is no easy way to determine which are more accurate. Furthermore, having alternative-splicing events exacerbates such difficult assembly problems. While benchmarking transcriptome assemblies is critical, this is also not trivial due to the general lack of true reference transcriptomes.

    Results

    In this study, we first provide a pipeline to generate a set of the simulated benchmark transcriptome and corresponding RNAseq data. Using the simulated benchmarking datasets, we compared the performance of various transcriptome assembly approaches including both de novo and genome-guided methods. The results showed that the assembly performance deteriorates significantly when alternative transcripts (isoforms) exist or for genome-guided methods when the reference is not available from the same genome. To improve the transcriptome assembly performance, leveraging the overlapping predictions between different assemblies, we present a new consensus-based ensemble transcriptome assembly approach, ConSemble.

    Conclusions

    Without using a reference genome, ConSemble using four de novo assemblers achieved an accuracy up to twice as high as any de novo assemblers we compared. When a reference genome is available, ConSemble using four genome-guided assemblies removed many incorrectly assembled contigs with minimal impact on correctly assembled contigs, achieving higher precision and accuracy than individual genome-guided methods. Furthermore, ConSemble using de novo assemblers matched or exceeded the best performing genome-guided assemblers even when the transcriptomes included isoforms. We thus demonstrated that the ConSemble consensus strategy both for de novo and genome-guided assemblers can improve transcriptome assembly. The RNAseq simulation pipeline, the benchmark transcriptome datasets, and the script to perform the ConSemble assembly are all freely available from:http://bioinfolab.unl.edu/emlab/consemble/.

     
    more » « less
  3. 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. 
    more » « less
  4. Abstract

    Transcriptome quality control is an important step in RNA‐Seq experiments. However, the quality of de novo assembled transcriptomes is difficult to assess, due to the lack of reference genome to compare the assembly to. We developed a method to assess and improve the quality of de novo assembled transcriptomes by focusing on the removal of chimeric sequences. These chimeric sequences can be the result of faulty assembled contigs, merging two transcripts into one. The developed method is incorporated into a pipeline, which we named Bellerophon, that is broadly applicable and easy to use. Bellerophon first uses the quality assessment tool TransRate to indicate the quality, after which it uses a transcripts per million (TPM) filter to remove lowly expressed contigs and CD‐HIT‐EST to remove highly identical contigs. To validate the quality of this method, we performed three benchmark experiments: (1) a computational creation of chimeras, (2) identification of chimeric contigs in a transcriptome assembly, (3) a simulated RNA‐Seq experiment using a known reference transcriptome. Overall, the Bellerophon pipeline was able to remove between 40% and 91.9% of the chimeras in transcriptome assemblies and removed more chimeric than nonchimeric contigs. Thus, the Bellerophon sequence of filtration steps is a broadly applicable solution to improve transcriptome assemblies.

     
    more » « less
  5. Abstract Motivation

    Accurate modeling of protein–protein interaction interface is essential for high-quality protein complex structure prediction. Existing approaches for estimating the quality of a predicted protein complex structural model utilize only the physicochemical properties or energetic contributions of the interacting atoms, ignoring evolutionarily information or inter-atomic multimeric geometries, including interaction distance and orientations.

    Results

    Here, we present PIQLE, a deep graph learning method for protein–protein interface quality estimation. PIQLE leverages multimeric interaction geometries and evolutionarily information along with sequence- and structure-derived features to estimate the quality of individual interactions between the interfacial residues using a multi-head graph attention network and then probabilistically combines the estimated quality for scoring the overall interface. Experimental results show that PIQLE consistently outperforms existing state-of-the-art methods including DProQA, TRScore, GNN-DOVE and DOVE on multiple independent test datasets across a wide range of evaluation metrics. Our ablation study and comparison with the self-assessment module of AlphaFold-Multimer repurposed for protein complex scoring reveal that the performance gains are connected to the effectiveness of the multi-head graph attention network in leveraging multimeric interaction geometries and evolutionary information along with other sequence- and structure-derived features adopted in PIQLE.

    Availability and implementation

    An open-source software implementation of PIQLE is freely available at https://github.com/Bhattacharya-Lab/PIQLE.

    Supplementary information

    Supplementary data are available at Bioinformatics Advances online.

     
    more » « less