nf-core/taxprofiler

Edit

Introduction

This document describes the output produced by the pipeline. Most of the plots are taken from the MultiQC report, which summarises results at the end of the pipeline.

The directories listed below will be created in the results directory after the pipeline has finished. All paths are relative to the top-level results directory.

Pipeline overview

The pipeline is built using Nextflow and processes data using the following steps:

• UNTAR - Optionally saved decompressed input databases
• FastQC - Raw read QC
• falco - Alternative to FastQC for raw read QC
• fastp - Adapter trimming for Illumina data
• AdapterRemoval - Adapter trimming for Illumina data
• Nonpareil - Read redundancy and metagenome coverage estimation for short reads
• BBDuk - Quality trimming and filtering for Illumina data
• PRINSEQ++ - Quality trimming and filtering for Illunina data
• Porechop - Adapter removal for Oxford Nanopore data
• Porechop_ABI - Adapter removal for Oxford Nanopore data
• Filtlong - Quality trimming and filtering for Nanopore data
• [Nanoq] (#nanoq) - Quality trimming and filtering for Nanopore data
• Bowtie2 - Host removal for Illumina reads
• minimap2 - Host removal for Nanopore reads
• SAMtools stats - Statistics from host removal
• SAMtools fastq - Converts unmapped BAM file to fastq format (minimap2 only)
• Analysis Ready Reads - Optional results directory containing the final processed reads used as input for classification/profiling.
• Bracken - Taxonomic classifier using k-mers and abundance estimations
• Kraken2 - Taxonomic classifier using exact k-mer matches
• KrakenUniq - Taxonomic classifier that combines the k-mer-based classification and the number of unique k-mers found in each species
• Centrifuge - Taxonomic classifier that uses a novel indexing scheme based on the Burrows-Wheeler transform (BWT) and the Ferragina-Manzini (FM) index.
• Kaiju - Taxonomic classifier that finds maximum (in-)exact matches on the protein-level.
• Diamond - Sequence aligner for protein and translated DNA searches.
• MALT - Sequence alignment and analysis tool designed for processing high-throughput sequencing data, especially in the context of metagenomics
• MetaPhlAn - Genome-level marker gene based taxonomic classifier
• mOTUs - Tool for marker gene-based OTU (mOTU) profiling.
• KMCP - Taxonomic classifier that utilizes genome coverage information by splitting the reference genomes into chunks and stores k-mers in a modified and optimized COBS index for fast alignment-free sequence searching.
• ganon - Taxonomic classifier and profile that uses Interleaved Bloom Filters as indices based on k-mers/minimizers.
• TAXPASTA - Tool to standardise taxonomic profiles as well as merge profiles across samples from the same database and classifier/profiler.
• MultiQC - Aggregate report describing results and QC from the whole pipeline
• Pipeline information - Report metrics generated during the workflow execution

untar

untar is used in nf-core/taxprofiler to decompress various input files ending in .tar.gz. This process is mainly used for decompressing input database archive files.

This directory will only be present if --save_untarred_databases is supplied. The contained directories can be useful for moving the decompressed directories to a central ‘cache’ location allowing users to re-use the same databases. This is useful to save unnecessary computational time of decompressing the archives on every run.

FastQC or Falco

FastQC gives general quality metrics about your sequenced reads. It provides information about the quality score distribution across your reads, per base sequence content (%A/T/G/C), adapter contamination and overrepresented sequences. For further reading and documentation see the FastQC help pages.

If preprocessing is turned on, nf-core/taxprofiler runs FastQC/Falco twice -once before and once after adapter removal/read merging, to allow evaluation of the performance of these preprocessing steps. Note in the General Stats table, the columns of these two instances of FastQC/Falco are placed next to each other to make it easier to evaluate. However, the columns of the actual preprocessing steps (i.e, fastp, AdapterRemoval, and Porechop) will be displayed after the two FastQC/Falco columns, even if they were run ‘between’ the two FastQC/Falco jobs in the pipeline itself.

fastp

fastp is a FASTQ pre-processing tool for quality control, trimmming of adapters, quality filtering and other features.

It is used in nf-core/taxprofiler for adapter trimming of short-reads.

By default nf-core/taxprofiler will only provide the <sample_id>.fastp.fastq.gz file if fastp is selected. The file <sample_id>.merged.fastq.gz will be available in the output folder if you provide the argument --shortread_qc_mergepairs (optionally retaining un-merged pairs when in combination with --shortread_qc_includeunmerged).

You can change the default value for low complexity filtering by using the argument --shortread_complexityfilter_fastp_threshold.

AdapterRemoval

AdapterRemoval searches for and removes remnant adapter sequences from High-Throughput Sequencing (HTS) data and (optionally) trims low quality bases from the 3’ end of reads following adapter removal. It is popular in the field of palaeogenomics. The output logs are stored in the results folder, and as a part of the MultiQC report.

By default nf-core/taxprofiler will only provide the .settings file if AdapterRemoval is selected.

You will only find the .fastq files in the results directory if you provide --save_preprocessed_reads. If this is selected, you may receive different combinations of .fastq files for each sample depending on the input types - e.g. whether you have merged or not, or if you’re supplying both single- and paired-end reads. Alternatively, if you wish only to have the ‘final’ reads that go into classification/profiling (i.e., that may have additional processing), do not specify this flag but rather specify --save_analysis_ready_reads, in which case the reads will be in the folder analysis_ready_reads.

Nonpareil

nonpareil is a tool for estimating metagenome ‘coverage’ from short-read libraries, i.e, whether all genomes within the metagenome have had at least one read sequenced. The lower the redundancy, the more sequencing can be done until the entire metagenome has been captured. The output can be used to guide the amount of further sequencing is required.

In most cases you will just want to look at the PNG files which contain the extrapolation information for estimating how much of the metagenome ‘coverage’ you will recover if you sequence more (i.e., to help indicate at what point you will just keep sequencing redundant reads that provide no more new taxonomic information).

The .npo files can be used for re-generating and customising the plots using the companion Nonpareil R package.

Porechop

Porechop is a tool for finding and removing adapters from Oxford Nanopore reads. Adapters on the ends of reads are trimmed and if a read has an adapter in its middle, it is considered a chimeric and it chopped into separate reads.

The output logs are saved in the output folder and are part of MultiQC report.You do not normally need to check these manually.

You will only find the .fastq files in the results directory if you provide --save_preprocessed_reads. Alternatively, if you wish only to have the ‘final’ reads that go into classification/profiling (i.e., that may have additional processing), do not specify this flag but rather specify --save_analysis_ready_reads, in which case the reads will be in the folder analysis_ready_reads.

Porechop_ABI

Porechop_ABI is an extension of Porechop. Unlike Porechop, Porechop_ABI does not use any external knowledge or database for the adapters. Adapters are discovered directly from the reads using approximate k-mers counting and assembly. Then these sequences can be used for trimming, using all standard Porechop options. The software is able to report a combination of distinct sequences if a mix of adapters is used. It can also be used to check whether a dataset has already been trimmed out or not, or to find leftover adapters in datasets that have been previously processed with Guppy.

The output logs are saved in the output folder and are part of MultiQC report.You do not normally need to check these manually.

You will only find the .fastq files in the results directory if you provide --save_preprocessed_reads. Alternatively, if you wish only to have the ‘final’ reads that go into classification/profiling (i.e., that may have additional processing), do not specify this flag but rather specify --save_analysis_ready_reads, in which case the reads will be in the folder analysis_ready_reads.

BBDuk

BBDuk stands for Decontamination Using Kmers. BBDuk was developed to combine most common data-quality-related trimming, filtering, and masking operations into a single high-performance tool.

It is used in nf-core/taxprofiler for complexity filtering using different algorithms. This means that it will remove reads with low sequence diversity (e.g. mono- or dinucleotide repeats).

By default nf-core/taxprofiler will only provide the .log file if BBDuk is selected as the complexity filtering tool. You will only find the complexity filtered reads in your results directory if you provide --save_complexityfiltered_reads. Alternatively, if you wish only to have the ‘final’ reads that go into classification/profiling (i.e., that may have additional processing), do not specify this flag but rather specify --save_analysis_ready_reads, in which case the reads will be in the folder analysis_ready_reads.

PRINSEQ++

PRINSEQ++ is a C++ implementation of the prinseq-lite.pl program. It can be used to filter, reformat or trim genomic and metagenomic sequence data.

It is used in nf-core/taxprofiler for complexity filtering using different algorithms. This means that it will remove reads with low sequence diversity (e.g. mono- or dinucleotide repeats).

By default nf-core/taxprofiler will only provide the .log file if PRINSEQ++ is selected as the complexity filtering tool. You will only find the complexity filtered .fastq files in your results directory if you supply --save_complexityfiltered_reads. Alternatively, if you wish only to have the ‘final’ reads that go into classification/profiling (i.e., that may have additional processing), do not specify this flag but rather specify --save_analysis_ready_reads, in which case the reads will be in the folder analysis_ready_reads.

Filtlong

Filtlong is a quality filtering tool for long reads. It can take a set of small reads and produce a smaller, better subset.

You will only find the .fastq files in the results directory if you provide --save_preprocessed_reads. Alternatively, if you wish only to have the ‘final’ reads that go into classification/profiling (i.e., that may have additional processing), do not specify this flag but rather specify --save_analysis_ready_reads, in which case the reads will be in the folder analysis_ready_reads.

Nanoq

nanoq is an ultra-fast quality filtering tool that also provides summary reports for nanopore reads.

You will only find the .fastq files in the results directory if you provide --save_preprocessed_reads. Alternatively, if you wish only to have the ‘final’ reads that go into classification/profiling (i.e., that may have additional processing), do not specify this flag but rather specify --save_analysis_ready_reads, in which case the reads will be in the folder analysis_ready_reads.

Bowtie2

Bowtie 2 is an ultrafast and memory-efficient tool for aligning sequencing reads to long reference sequences. It is particularly good at aligning reads of about 50 up to 100s or 1,000s of characters, and particularly good at aligning to relatively long (e.g. mammalian) genomes.

It is used with nf-core/taxprofiler to allow removal of ‘host’ (e.g. human) and/or other possible contaminant reads (e.g. Phi X) from short-read .fastq files prior to profiling.

By default nf-core/taxprofiler will only provide the .log file if host removal is turned on. You will only have a .bam file if you specify --save_hostremoval_bam. This will contain both mapped and unmapped reads. You will only get FASTQ files if you specify to save --save_hostremoval_unmapped - these contain only unmapped reads. Alternatively, if you wish only to have the ‘final’ reads that go into classification/profiling (i.e., that may have additional processing), do not specify this flag but rather specify --save_analysis_ready_reads, in which case the reads will be in the folder analysis_ready_reads.

minimap2

minimap2 is an alignment tool suited to mapping long reads to reference sequences.

It is used with nf-core/taxprofiler to allow removal of ‘host’ (e.g. human) or other possible contaminant reads from long-read .fastq files prior to taxonomic classification/profiling.

By default, nf-core/taxprofiler will only provide the .bam file containing mapped and unmapped reads if saving of host removal for long reads is turned on via --save_hostremoval_bam.

SAMtools fastq

SAMtools fastq converts a .sam, .bam, or .cram alignment file to FASTQ format

This directory will be present and contain the unmapped reads from the .fastq format from long-read minimap2 host removal, if --save_hostremoval_unmapped is supplied. Alternatively, if you wish only to have the ‘final’ reads that go into classification/profiling (i.e., that may have additional processing), do not specify this flag but rather specify --save_analysis_ready_reads, in which case the reads will be in the folder analysis_ready_reads.

Analysis Ready Reads

The results directory will contain the ‘final’ processed reads used as input for classification/profiling. It will only include the output of the last step of any combinations of preprocessing steps that may have been specified in the run configuration. For example, if you perform the read QC and host-removal preprocessing steps, the final reads that are sent to classification/profiling are the host-removed FASTQ files - those will be the ones present in this directory.

SAMtools stats

SAMtools stats collects statistics from a .sam, .bam, or .cram alignment file and outputs in a text format.

In most cases you do not need to check this file, as it is rendered in the MultiQC run report.

Run Merging

nf-core/taxprofiler offers the option to merge FASTQ files of multiple sequencing runs or libraries that derive from the same sample, as specified in the input samplesheet.

This is the last possible preprocessing step, so if you have multiple runs or libraries (and run merging turned on), this will represent the final reads that will go into classification/profiling steps.

Note that you will only find samples that went through the run merging step in this directory. For samples that had a single run or library will not go through this step of the pipeline and thus will not be present in this directory.

This directory and its FASTQ files will only be present if you supply --save_runmerged_reads.Alternatively, if you wish only to have the ‘final’ reads that go into classification/profiling (i.e., that may have additional processing), do not specify this flag but rather specify --save_analysis_ready_reads, in which case the reads will be in the folder analysis_ready_reads.

Bracken

Bracken (Bayesian Reestimation of Abundance with Kraken) is a highly accurate statistical method that computes the abundance of species in DNA sequences from a metagenomics sample. Braken uses the taxonomy labels assigned by Kraken, a highly accurate metagenomics classification algorithm, to estimate the number of reads originating from each species present in a sample.

The main taxonomic profiling file from Bracken is the *.tsv file. This provides the basic results from Kraken2 but with the corrected abundance information. Note that the raw Kraken2 version of the upstream step of Bracken can be found in the kraken2/ directory with the suffix of <sample_id>_<db_name>.bracken.report.txt (with a 6 column variant when --save_minimizers specified).

Kraken2

Kraken is a taxonomic sequence classifier that assigns taxonomic labels to DNA sequences. Kraken examines the k-mers within a query sequence and uses the information within those k-mers to query a database. That database maps -mers to the lowest common ancestor (LCA) of all genomes known to contain a given k-mer.

The main taxonomic classification file from Kraken2 is the _combined_reports.txt or *report.txt file. The former provides you the broadest over view of the taxonomic classification results across all samples against a single database, where you get two columns for each sample e.g. 2_all and 2_lvl, as well as a summarised column summing up across all samples tot_all and tot_lvl. The latter gives you the most information for a single sample. The report file is also used for the taxpasta step.

You will only receive the .fastq and *classifiedreads.txt file if you supply --kraken2_save_reads and/or --kraken2_save_readclassifications parameters to the pipeline.

When running Bracken, you will only get the ‘intermediate’ Kraken2 report files in this directory if you supply --bracken_save_intermediatekraken2 to the run.

KrakenUniq

KrakenUniq (formerly KrakenHLL) is an extension to the fast k-mer-based classification performed by Kraken with an efficient algorithm for additionally assessing the coverage of unique k-mers found in each species in a dataset.

The main taxonomic classification file from KrakenUniq is the *.krakenuniq.report.txt file. This is an extension of the Kraken2 report with the additional k-mer coverage information that provides more information about the accuracy of hits.

You will only receive the .fasta.gz and *.krakenuniq.classified.txt file if you supply --krakenuniq_save_reads and/or --krakenuniq_save_readclassification parameters to the pipeline.

Centrifuge

Centrifuge is a taxonomic sequence classifier that uses a Burrows-Wheeler transform and Ferragina-Manzina index for storing and mapping sequences.

The main taxonomic classification files from Centrifuge are the _combined_reports.txt, *report.txt, *results.txt and the *centrifuge.txt. The latter is used by the taxpasta step. You will receive the .fastq files if you supply --centrifuge_save_reads.

Kaiju

Kaiju is a taxonomic classifier that finds maximum exact matches on the protein-level using the Burrows-Wheeler transform.

The most useful summary file is the _combined_reports.txt file which summarises hits across all reads and samples. Separate per-sample versions summaries can be seen in <db>/*.txt. However if you wish to look at more precise information on a per-read basis, see the *tsv file. The default taxonomic rank is species. You can provide a different one by updating the argument --kaiju_taxon_rank.

DIAMOND

DIAMOND is a sequence aligner for translated DNA searches or protein sequences against a protein reference database such as NR. It is a replacement for the NCBI BLAST software tools.It has many key features and it is used as taxonomic classifier in nf-core/taxprofiler.

By default you will receive a TSV output. Alternatively, you will receive a *.sam file if you provide the parameter --diamond_save_reads but in this case no taxonomic classification will be available(!), only the aligned reads in sam format.

MALT

MALT is a fast replacement for BLASTX, BLASTP and BLASTN, and provides both local and semi-global alignment capabilities.

The main output of MALT is the .rma6 file format, which can be only loaded into MEGAN and it’s related tools. We provide the rma2info text files for improved compatibility with spreadsheet programs and other programmtic data manipulation tools, however this has only limited information compared to the ‘binary’ RMA6 file format (the .txt file only contains taxonomic ID and count, whereas RMA6 has taxonomic lineage information).

You will only receive the .sam and .megan files if you supply --malt_save_reads and/or --malt_generate_megansummary parameters to the pipeline.

MetaPhlAn

MetaPhlAn is a computational tool for profiling the composition of microbial communities (Bacteria, Archaea and Eukaryotes) from metagenomic shotgun sequencing data (i.e. not 16S) with species-level resolution via marker genes.

The output contains a file named *_combined_reports.txt, which provides an overview of the classification results for all samples. The main taxonomic profiling file from MetaPhlAn is the *_profile.txt file. This provides the abundance estimates from MetaPhlAn however does not include raw counts by default. Additionally, it contains intermediate Bowtie2 output .bowtie2out.txt, which presents a condensed representation of the mapping results of your sequencing reads to MetaPhlAn’s marker gene sequences. The alignments are listed in tab-separated columns, including Read ID and Marker Gene ID, with each alignment represented on a separate line.

mOTUs

mOTUS is a taxonomic profiler that maps reads to a unique marker specific database and estimates the relative abundance of known and unknown species.

Normally *_combined_reports.txt is the most useful file for downstream analyses, but the per sample .out file can provide additional more specific information. By default, nf-core/taxprofiler is providing a column describing NCBI taxonomic ID as this is used in the taxpasta step. You can disable this column by activating the argument --motus_remove_ncbi_ids. You will receive the relative abundance instead of read counts if you provide the argument --motus_use_relative_abundance.

KMCP

KMCP utilises genome coverage information by splitting the reference genomes into chunks and stores k-mers in a modified and optimised COBS index for fast alignment-free sequence searching. KMCP combines k-mer similarity and genome coverage information to reduce the false positive rate of k-mer-based taxonomic classification and profiling methods.

You will receive the <sample_id>.gz file if you supply --kmcp_save_search. Please note that there is no taxonomic label assignment in this output file.

The main taxonomic classification file from KMCP is the *kmcp.profile which is also used by the taxpasta step.

ganon

ganon is designed to index large sets of genomic reference sequences and to classify reads against them efficiently. The tool uses Interleaved Bloom Filters as indices based on k-mers/minimizers. It was mainly developed, but not limited, to the metagenomics classification problem: quickly assign sequence fragments to their closest reference among thousands of references. After classification, taxonomic abundance is estimated and reported.

Generally you will want to refer to the combined_reports.txt or _report.tre file. For further descriptions of the contents of each file, see the ganon documentation.

You will only receive the .all, .lca, and .unc files if you supply the --ganon_save_readclassifications parameter to the pipeline.

Krona

Krona allows the exploration of (metagenomic) hierarchical data with interactive zooming, multi-layered pie charts.

Krona charts will be generated by the pipeline for supported tools (Kraken2, Centrifuge, Kaiju, and MALT)

The resulting HTML files can be loaded into your web browser for exploration. Each file will have a dropdown to allow you to switch between each sample aligned against the given database of the tool.

TAXPASTA

TAXPASTA standardises and optionally merges two or more taxonomic profiles across samples into one single table. It supports multiple different classifiers simplifying comparison of taxonomic classification results between tools and databases.

By providing the path to a directory containing taxdump files to --taxpasta_taxonomy_dir, the taxon name, the taxon rank, the taxon’s entire lineage including taxon names and/or the taxon’s entire lineage including taxon identifiers can also be added in the output in addition to just the taxon ID. Addition of this extra information can be turned by using the parameters --taxpasta_add_name, --taxpasta_add_rank, --taxpasta_add_lineage and --taxpasta_add_idlineage respectively.

These files will likely be the most useful files for the comparison of differences in classification between different tools or building consensuses, with the caveat they have slightly less information than the actual output from each tool (which may have non-standard information e.g. taxonomic rank, percentage of hits, abundance estimations).

The following report files are used for the taxpasta step:

• Bracken: <sample>_<db_name>.tsv Taxpasta used the new_est_reads column for the standardised profile.
• Centrifuge: <sample_id>.centrifuge.txt Taxpasta uses the direct_assigned_reads column for the standardised profile.
• Diamond: <sample_id> Taxpasta summarises number of reads per NCBI taxonomy ID standardised profile.
• Kaiju: <sample_id>_<db_name>.kaijutable.txt Taxpasta uses the reads column from kaiju2table standardised profile.
• KrakenUniq: <sample_id>_<db_name>.report.txt Taxpasta uses the reads column for the standardised profile.
• Kraken2: <sample_id>_<db_name>.report.txt Taxpasta uses the direct_assigned_reads column for the standardised profile.
• MALT: <sample_id>.txt.gz Taxpasta uses the count (second) column from the output of MEGAN6’s rma2info for the standardised profile.
• MetaPhlAn: <sample_id>_profile.txt Taxpasta uses the relative_abundance column multiplied with a fixed number to yield an integer for the standardised profile.
• mOTUs: <sample_id>.out Taxpasta uses the read_count column for the standardised profile.

MultiQC

MultiQC is a visualization tool that generates a single HTML report summarising all samples in your project. Most of the pipeline QC results are visualised in the report and further statistics are available in the report data directory.

Results generated by MultiQC collate pipeline QC from supported tools e.g. FastQC. The pipeline has special steps which also allow the software versions to be reported in the MultiQC output for future traceability. For more information about how to use MultiQC reports, see http://multiqc.info.

All tools in taxprofiler supported by MultiQC will have a dedicated section showing summary statistics of each tool based on information stored in log files.

You can expect in the MultiQC reports either sections and/or general stats columns for the following tools:

• fastqc
• adapterremoval
• fastp
• nonpareil
• bbduk
• prinseqplusplus
• nanoq
• porechop_abi
• porechop
• filtlong
• bowtie2
• minimap2
• samtools (stats)
• kraken
• bracken
• centrifuge
• kaiju
• diamond
• malt
• motus

Pipeline information

Nextflow provides excellent functionality for generating various reports relevant to the running and execution of the pipeline. This will allow you to troubleshoot errors with the running of the pipeline, and also provide you with other information such as launch commands, run times and resource usage.