Critical Assessment of Metagenome Interpretation—a benchmark of metagenomics software

Abstract

In metagenome analysis, computational methods for assembly, taxonomic profiling and binning are key components facilitating downstream biological data interpretation. However, a lack of consensus about benchmarking datasets and evaluation metrics complicates proper performance assessment. The Critical Assessment of Metagenome Interpretation (CAMI) challenge has engaged the global developer community to benchmark their programs on datasets of unprecedented complexity and realism. Benchmark metagenomes were generated from ~700 newly sequenced microorganisms and ~600 novel viruses and plasmids, including genomes with varying degrees of relatedness to each other and to publicly available ones and representing common experimental setups. Across all datasets, assembly and genome binning programs performed well for species represented by individual genomes, while performance was substantially affected by the presence of related strains. Taxonomic profiling and binning programs were proficient at high taxonomic ranks, with a notable performance decrease below the family level. Parameter settings substantially impacted performances, underscoring the importance of program reproducibility. While highlighting current challenges in computational metagenomics, the CAMI results provide a roadmap for software selection to answer specific research questions.

Towards reproducible, transparent, and systematic benchmarking of omics computational tools

“…Further, 50% of surveyed benchmarking studies do not to share supporting documentation on installing and running the benchmark tools (Table 1). Ideally, supporting documentation needs to be shared as an easyto-use interface instead of main text and/or supplementary materials, both of which make it easier for researchers to adopt the method recommended by the benchmarking study 8,56 .…”

“…Prepare a 56 , an easy-to-use interface that packages software with all its dependencies and parameters, enabling flexibility and portability of the software tool across platforms and operating systems. To the best of our knowledge, only a single benchmarking study (CAMI) 8 wrapped tools as portable containers. Prior to running the software, contact the software developers to ensure the correctness of chosen commands and parameters.…”

“…Critical Assessment of Metagenome Interpretation (CAMI) 8,37 and is available via GitHub (https://github.com/dkoslicki/CAMIProfilingAnalysis). Ideally, the repositories storing code and data should be permanently archived (for example, via Zenodo https://zenodo.org/) to prevent stored materials from being changed or removed.…”

“…List of names of the studies with corresponding citation used in Table 1. • A survey of error-correction methods for next-generation sequencing 40 • Critical assessment of automated flow cytometry data analysis techniques 19 • Assemblathon 2: evaluating de novo methods of genome assembly in three vertebrate species 41 • A comprehensive evaluation of assembly scaffolding tools 37 • An evaluation of the accuracy and speed of metagenome analysis tools 27,37 • Critical Assessment of Metagenome Interpretation-a benchmark of metagenomics software 8 • Comprehensive benchmarking and ensemble approaches for metagenomic classifiers 34 • A large-scale evaluation of computational protein function prediction 42 • An expanded evaluation of protein function prediction methods shows an improvement in accuracy 43 • Standardized benchmarking in the quest for orthologs 18 • Benchmarking short sequence mapping tools 44 • Alignathon: a competitive assessment of whole-genome alignment methods 45 • Simulation-based comprehensive benchmarking of RNA-seq aligners 6 • Characterization and improvement of RNA-Seq precision in quantitative transcript expression profiling 6,46 • Assessment of transcript reconstruction methods for RNA-seq 47 …”

“…The developer-researcher gap can be addressed with benchmarking studies, which inform the research community about the most appropriate tools for specific analytical tasks and data types [6][7][8] . A benchmarking study consists of a robust and comprehensive evaluation of the abilities of existing algorithms to solve a particular computational biology problem.…”