DADA2 Pipeline

R packages

Nephele runs the DADA2 R package v1.10 following the steps in the package authors' Big Data workflow including optional use of DECIPHER package v2.10. We make some minor modifications of the parameters used. Our pipeline is outlined below. If you are new to DADA2, it might be helpful to read through the DADA2 Tutorial.

User Options

  • Ion Torrent Data - Beta: By default, DADA2 is trained to work on Illumina data. Checking this option sets the denoising parameters according to DADA2's suggested values for Ion Torrent data. They also suggest the trim left parameter be increased by 15 bp (on top of any primer lengths). This option is in beta, and has not been extensively tested. If you have Ion Torrent data, we are interested in your feedback - please email us!

Filter and Trim

  • Trim left: The number of nucleotides to remove from the start of each read, forward and reverse. The values should be chosen based on the lengths of primers used for sequencing. If your data are untrimmed, this parameter is very important for the DADA2 pipeline. See this FAQ (Default: 20).
  • Truncation quality score: Truncate reads at the first instance of a quality score less than or equal to this value. (Default: 4).
  • Truncation length: The length at which to truncate reads, forward and reverse. Reads shorter than these lengths are discarded. If set to 0, reads are not truncated. If both trim left and truncation length are set, the filtered reads will have length = truncation length - trim left. (Default: 0).
  • Maximum expected errors (maxEE): After truncation, reads with higher than this many "expected errors" will be discarded. Expected errors are calculated from the nominal definition of the quality score: EE = sum(10^(-Q/10)). (Default: 5).

Merge Pairs

For paired-end data only.

  • Just concatenate: Concatenate paired reads instead of merging. (Default: FALSE)
  • Maximum mismatches: The maximum number of mismatches allowed in the overlap region when merging read pairs. (Default: 0).
  • Trim overhanging sequence: After merging paired end reads, trim sequence which overhangs the start of each read. If amplicons are shorter than read length, e.g. 16S V4 region, we suggest checking this option. (Logical. Default: FALSE).

Analysis

  • Chimera removal: Remove chimeric sequences. If primers are not trimmed (either prior to submission or using the trim left option), then we suggest unchecking this option. (Default: True).
  • Taxonomic assigment: Method to be used for taxonomic assignment, either rdp or IDTAXA. (Default: rdp)
  • Reference database: Reference database to be used for taxonomic assignment. IDTAXA will use its own SILVA v132 database. See Databases below.
  • Sampling depth: The number of counts for filtering and subsampling the OTU table for downstream analysis. Samples which have counts below this value will be removed from the downstream analysis. The counts of the remaining samples will be subsampled to this value. If not specified, it will be calculated automatically. (See this FAQ)

Databases

Pipeline steps

  1. Plot quality profiles of forward and reverse reads. These graphs are saved as qualityProfile_R1.pdf and qualityProfile_R2.pdf.

    pqp1 <- plotQualityProfile(file.path(datadir, r1))
    pqp2 <- plotQualityProfile(file.path(datadir, r2))
  2. Preprocess sequence data with filterAndTrim. The maxEE, trimLeft, truncQ, and truncLen parameters can be set by the user (defaults used below as example). The filtered sequence files, *_trim.fastq.gz, are output to the filtered_data directory.

    filterAndTrim(fwd = file.path(datadir, readslist$R1), filt = file.path(filt.dir, trimlist$R1), 
        rev = file.path(datadir, readslist$R2), filt.rev = file.path(filt.dir, trimlist$R2), 
        maxEE = 5, trimLeft = 20, truncQ = 4, truncLen = 0, rm.phix = TRUE, compress = TRUE, 
        verbose = TRUE, multithread = nthread, minLen = 50)
  3. Learn the error rates with learnErrors. The nbases parameter is set to 1e+08. The error rate graphs made with plotErrors are saved as errorRate_R1.pdf, errorRate_R2.pdf. The error profiles, err, are also saved as a list R binary object in the intermediate_files directory.

    errR1 <- learnErrors(r1, multithread = nthread, nbases = nbases, randomize = TRUE)
    pe1 <- plotErrors(errR1, nominalQ = TRUE)
  4. Dereplicate reads with derepFastq and run the dada sequence-variant inference algorithm.

    derepR1 <- derepFastq(r1[sample], verbose = TRUE)
    ddR1 <- dada(derepR1, err = errR1, multithread = nthread, verbose = 1)
  5. For paired-end data, merge the overlapping denoised reads with mergePairs. The minOverlap parameter is set to 12. trimOverhang, justConcatenate, and maxMismatch are set by the user. The sequence table, seqtab, containing the final amplicon sequence variants (ASVs), is saved as an R binary object to the intermediate_files directory.

    mergePairs(dd$R1, derep$R1, dd$R2, derep$R2, verbose = TRUE, minOverlap = 12, trimOverhang = FALSE, 
        justConcatenate = FALSE, maxMismatch = 0)
    seqtab <- makeSequenceTable(mergedReads)
  6. Filter out ASVs of length less than 75 bp. The sequence table is saved as seqtab_min75.rds. Also, filter out chimeras with removeBimeraDenovo, if the option is chosen.

    seqtab <- seqtab[, which(seqlengths >= 75)]
    seqtabnochimera <- removeBimeraDenovo(seqtab, verbose = TRUE, multithread = nthread)
  7. Classify the remaining ASVs taxonomically with

    • rdp using assignTaxonomy (default). The minBoot parameter for minimum bootstrap confidence is set to 80 and tryRC is set to TRUE. This genus level result is saved as taxa.biom. Add species annotation to the taxonomic identification using addSpecies. This final result is saved as a biom file taxa_species.biom.

      taxa <- assignTaxonomy(seqtab, refdb, multithread = nthread, minBoot = 80, tryRC = TRUE, 
          verbose = TRUE)
      taxa.species <- addSpecies(taxa, refdb_species, verbose = TRUE, tryRC = TRUE)

      For PE data, if the mergePairs justConcatenate option is checked, species annotation will only be done using the forward reads (R1).

    • IDTAXA using IdTaxa from the DECIPHER R package. The final result will be saved as taxa.biom

      dna <- DNAStringSet(getSequences(seqtab))
      ids <- IdTaxa(dna, trainingSet, strand = "both", processors = nthread, verbose = T)
  8. The final results are also saved as a tab-separated text file OTU_table.txt. The final sequence variants used for taxonomic classification are output as seq.fasta. A summary of the counts in the OTU table is saved to otu_summary_table.txt.

Output Files

Complete descriptions of the intermediate and final output files can be found in the Pipeline Steps above.

  • OTU_table.txt: tab-separated text file of ASV counts and taxonomic assignment
  • seq.fasta: FASTA file of amplicon sequence variants
  • taxa_species.biom: taxonomic assignment at the species level BIOM V1 format
  • otu_summary_table.txt: summary of the sequence variant counts by sample
  • taxa.biom: BIOM V1 file of sequence variant counts with taxonomic assignment at the genus level
  • taxonomy_table.txt: tab-separated taxonomy file suitable for importing into QIIME2
  • errorRate_R1/2.pdf: error profile plots
  • qualityProfile_R1/2.pdf: quality profile plots
  • filtered_data: trimmed sequence files
  • intermediate_files: intermediate files produced by the pipeline; useful for debugging
  • graphs: output of the visualization pipeline

Tools and References

  1. Callahan BJ, McMurdie PJ, Rosen MJ, Han AW, Johnson AJA and Holmes SP (2016). "DADA2: High-resolution sample inference from Illumina amplicon data." Nature Methods, 13, pp. 581-583. doi: 10.1038/nmeth.3869.

  2. McMurdie PJ and Paulson JN (2016). biomformat: An interface package for the BIOM file format. https://github.com/joey711/biomformat/.

  3. Microsoft and Weston S (2017). foreach: Provides Foreach Looping Construct for R. R package version 1.4.4, https://CRAN.R-project.org/package=foreach.

  4. Murali, A., Bhargava, A., and Wright, E. S. (2018). IDTAXA: a novel approach for accurate taxonomic classification of microbiome sequences. Microbiome, 6(1). doi: 10.1186/s40168-018-0521-5.

  5. Quast C., Pruesse E., Yilmaz P., Gerken, J., Schweer T., Yarza P., Peplies, J., Glöckner, F. O. (2013). "The SILVA ribosomal RNA gene database project: Improved data processing and web-based tools." Nucleic Acids Research, 41(D1), D590-D596. doi: 10.1093/nar/gks1219.

  6. Escapa, I. F., Chen, T., Huang, Y., Gajare, P., Dewhirst, F. E., & Lemon, K. P. (2018). "New Insights into Human Nostril Microbiome from the Expanded Human Oral Microbiome Database (eHOMD): a Resource for the Microbiome of the Human Aerodigestive Tract." MSystems, 3(6), e00187-18. doi: 10.1128/mSystems.00187-18.

  7. DeSantis, T. Z., et al. "Greengenes, a Chimera-Checked 16S rRNA Gene Database and Workbench Compatible with ARB." Applied and Environmental Microbiology, vol. 72, no. 7, July 2006, pp. 5069–72. aem.asm.org, doi: 10.1128/AEM.03006-05.