nf-core/ampliseq

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:

• nf-core/ampliseq: Output
  • Pipeline overview
    • FastQC - Read quality control
    • Cutadapt - Primer trimming
    • MultiQC - Aggregate report describing results
    • DADA2 - Infer Amplicon Sequence Variants (ASVs) and taxonomic classification
      • Barrnap - Predict ribosomal RNA sequences
      • ITSx - Optionally, taxonomic classification can be performed on ITS region only
    • QIIME2 - Secondary analysis
      • Taxonomic classification - Taxonomical classification of ASVs
      • Exclude taxa - Remove unwanted ASV based on taxonomy
      • Relative abundance tables - Exported relative abundance tables
      • Barplot - Interactive barplot
      • Alpha diversity rarefaction curves - Rarefaction curves for quality control
      • Diversity analysis - High level overview with different diversity indices
        • Alpha diversity indices - Diversity within samples
        • Beta diversity indices - Diversity between samples (e.g. PCoA plots)
      • ANCOM - Differential abundance analysis
    • PICRUSt2 - Predict the functional potential of a bacterial community
    • Read count report - Report of read counts during various steps of the pipeline
    • Pipeline information - Report metrics generated during the workflow execution
  • Citations

FastQC

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.

Cutadapt

Cutadapt is trimming primer sequences from sequencing reads. Primer sequences are non-biological sequences that often introduce point mutations that do not reflect sample sequences. This is especially true for degenerated PCR primer. If primer trimming would be omitted, artifactual amplicon sequence variants might be computed by the denoising tool or sequences might be lost due to become labelled as PCR chimera.

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.

NB: The FastQC plots displayed in the MultiQC report shows untrimmed reads. They may contain adapter sequence and potentially regions with low quality.

DADA2

DADA2 performs fast and accurate sample inference from amplicon data with single-nucleotide resolution. It infers exact amplicon sequence variants (ASVs) from amplicon data with fewer false positives than many other methods while maintaining high sensitivity.

DADA2 computes an error model on the sequencing reads (forward and reverse independently), therefore quality filtering or paired read merging may not be performed before. Each sequencing run varies in their error profile and it is recommended that DADA2 runs separately on data from each run individually. It is recommended to use the ampliseq option --multiple_sequencing_runs to analyse such data.

DADA2 reduces sequence errors and dereplicates sequences by quality filtering, denoising, read pair merging (for paired end Illumina reads only) and PCR chimera removal.

Additionally, DADA2 taxonomically classifies the ASVs using a choice of supplied databases (specified with --dada_ref_taxonomy).

Barrnap

Barrnap predicts the location of ribosomal RNA genes in genomes, here it can be used to discriminate rRNA sequences from potential contamination. It supports bacteria (5S,23S,16S), archaea (5S,5.8S,23S,16S), metazoan mitochondria (12S,16S) and eukaryotes (5S,5.8S,28S,18S).

Optionally, ASV sequences can be filtered for rRNA sequences identified by Barrnap with --filter_ssu that can take a list of abbreviations of the above supported categories (kingdoms), e.g. bac,arc,mito,euk. This filtering takes place after DADA2’s ASV computation (i.e. after chimera removal) but before taxonomic classification (also applies to above mentioned taxonomic classification with DADA2, i.e. files ASV_tax.tsv & ASV_tax_species.tsv).

ITSx

Optionally, the ITS region can be extracted from each ASV sequence using ITSx, and taxonomic classification is performed based on the ITS sequence.

QIIME2

Quantitative Insights Into Microbial Ecology 2 (QIIME2) is a next-generation microbiome bioinformatics platform and the successor of the widely used QIIME1.

ASV sequences, counts, and taxonomic classification as produced before with DADA2 are imported into QIIME2 and further analysed. Optionally, ASVs can be taxonomically classified also with QIIME2 against a database chosen with --qiime_ref_taxonomy (but DADA2 taxonomic classification takes precedence). Next, ASVs are filtered (--exclude_taxa, --min_frequency, --min_samples), and abundance tables are exported. Following, diversity indices are calculated and testing for differential abundant features between sample groups is performed.

Taxonomic classification

Taxonomic classification with QIIME2 is typically similar to DADA2 classifications. However, both options are available. When taxonomic classification with DADA2 and QIIME2 is performed, DADA2 classification takes precedence over QIIME2 classifications for all downstream analysis.

Exclude taxa

Removes unwanted taxa in DADA2 output sequences and abundance tables by taxonomic classification. Unwanted taxa are often off-targets generated in PCR with primers that are not perfectly specific for the target DNA. For example, PCR with commonly used primers also amplifyies mitrochindrial or chloroplast rRNA genes and therefore leads to non-bacteria products. These mitrochondria or chloroplast amplicons are removed in this step by default (--exclude_taxa). The tables are based on the computed taxonomic classification (DADA2 classification takes precedence over QIIME2 classifications).

All following analysis is based on these filtered tables.

Relative abundance tables

Absolute abundance tables produced by the previous steps contain count data, but the compositional nature of 16S rRNA amplicon sequencing requires sequencing depth normalisation. This step computes relative abundance tables for various taxonomic levels and detailed tables for all ASVs with taxonomic classification, sequence and relative abundance for each sample. Typically used for in depth investigation of taxa abundances. If not specified, the tables are based on the computed taxonomic classification (DADA2 classification takes precedence over QIIME2 classifications).

Barplot

Produces an interactive abundance plot count tables that aids exploratory browsing the discovered taxa and their abundance in samples and allows sorting for associated meta data, DADA2 classification takes precedence over QIIME2 classifications.

Alpha diversity rarefaction curves

Produces rarefaction plots for several alpha diversity indices, and is primarily used to determine if the richness of the samples has been fully observed or sequenced. If the slope of the curves does not level out and the lines do not becomes horizontal, this might be because the sequencing depth was too low to observe all diversity or that sequencing error artificially increases sequence diversity and causes false discoveries.

Diversity analysis

Diversity measures summarize important sample features (alpha diversity) or differences between samples (beta diversity). To do so, sample data is first rarefied to the minimum number of counts per sample. Also, a phylogenetic tree of all ASVs is computed to provide phylogenetic information.

Alpha diversity indices

Alpha diversity measures the species diversity within samples. Diversity calculations are based on sub-sampled data rarefied to the minimum read count of all samples. This step calculates alpha diversity using various methods and performs pairwise comparisons of groups of samples. It is based on a phylogenetic tree of all ASV sequences.

Beta diversity indices

Beta diversity measures the species community differences between samples. Diversity calculations are based on sub-sampled data rarefied to the minimum read count of all samples. This step calculates beta diversity distances using various methods and performs pairwise comparisons of groups of samples. Additionally, principle coordinates analysis (PCoA) plots are produced that can be visualized with Emperor in your default browser without the need for installation. This calculations are based on a phylogenetic tree of all ASV sequences. Furthermore, ADONIS permutation-based statistical test in vegan-R determine whether groups of samples are significantly different from one another. By default, all metadata columns that are for suitable pairwise comparisons will be tranformed into a formula, e.g. “treatment1+treatment2”. A custom formula can be supplied with --qiime_adonis_formula.

The following methods are used to calculate community dissimilarities:

• Jaccard distance (qualitative)
• Bray-Curtis distance (quantitative)
• unweighted UniFrac distance (qualitative, phylogenetic)
• weighted UniFrac distance (quantitative, phylogenetic)

ANCOM

Analysis of Composition of Microbiomes (ANCOM) is applied to identify features that are differentially abundant across sample groups. A key assumption made by ANCOM is that few taxa (less than about 25%) will be differentially abundant between groups otherwise the method will be inaccurate.

ANCOM is applied to each suitable or specified metadata column for 5 taxonomic levels (2-6).

PICRUSt2

PICRUSt2 (Phylogenetic Investigation of Communities by Reconstruction of Unobserved States) is a software for predicting functional abundances based only on marker gene sequences. On demand (--picrust), Enzyme Classification numbers (EC), KEGG orthologs (KO) and MetaCyc ontology predictions will be made for each sample.

PICRUSt2 is preferentially applied to filtered data by QIIME2 but will use DADA2 output in case QIIME2 isnt run.

NB: Quantifications are not normalized yet, they can be normalized e.g. by the total sum per sample.

SBDI export

You can use the --sbdiexport flag (or sbdiexport: true in a nextflow parameter file using -params-file in yml format) to generate tab separated files in preparation for submission to the Swedish Biodiversity Infrastructure (SBDI).

Tables are generated from the DADA2 denoising and taxonomy assignment steps. Each table, except annotation.tsv, corresponds to one tab in the submission template. See docs/usage.md for further information. Most of the fields in the template will not be populated by the export process, but if you run nf-core/ampliseq with a sample metadata table (--metadata) any fields corresponding to a field in the template will be used.

Read count report

This report includes information on how many reads per sample passed each pipeline step in which a loss can occur. Specifically, how many read pairs entered cutadapt, were reverse complemented, passed trimming; how many read pairs entered DADA2, were denoised, merged and non-chimeric; and how many counts were lost during excluding unwanted tax and removing low abundance/prevalence sequences in QIIME2.

Pipeline information

Citations

An extensive list of references for the tools used by the pipeline can be found in the CITATIONS.md file.