DADA2 is a software package that models and corrects Illumina-sequenced amplicon errors. DADA2 infers sample sequences exactly, without coarse-graining into OTUs, and resolves differences of as little as one nucleotide. DADA2 identified more real variants and output fewer spurious sequences than other methods.
DADA2’s advantage is that it uses more of the data. The DADA2 error model incorporates quality information, which is ignored by all other methods after filtering. The DADA2 error model incorporates quantitative abundances, whereas most other methods use abundance ranks if they use abundance at all. The DADA2 error model identifies the differences between sequences, eg. A->C, whereas other methods merely count the mismatches. DADA2 can parameterize its error model from the data itself, rather than relying on previous datasets that may or may not reflect the PCR and sequencing protocols used in your study.
DADA2 Software Package is available as an R package at : https://benjjneb.github.io/dada2/index.html
DADA2 pipeline includes several tools for read quality control, including quality filtering, trimming, denoising, pair merging and chimera filtering. Below are the major processing steps of DADA2:
Step 1. Read trimming based on sequence quality The quality of NGS Illumina sequences often decreases toward the end of the reads. DADA2 allows to trim off the poor quality read ends in order to improve the error model building and pair mergicing performance.
Step 2. Learn the Error Rates The DADA2 algorithm makes use of a parametric error model (err) and every amplicon dataset has a different set of error rates. The learnErrors method learns this error model from the data, by alternating estimation of the error rates and inference of sample composition until they converge on a jointly consistent solution. As in many machine-learning problems, the algorithm must begin with an initial guess, for which the maximum possible error rates in this data are used (the error rates if only the most abundant sequence is correct and all the rest are errors).
Step 3. Infer amplicon sequence variants (ASVs) based on the error model built in previous step. This step is also called sequence "denoising". The outcome of this step is a list of ASVs that are the equivalent of oligonucleotides.
Step 4. Merge paired reads. If the sequencing products are read pairs, DADA2 will merge the R1 and R2 ASVs into single sequences. Merging is performed by aligning the denoised forward reads with the reverse-complement of the corresponding denoised reverse reads, and then constructing the merged “contig” sequences. By default, merged sequences are only output if the forward and reverse reads overlap by at least 12 bases, and are identical to each other in the overlap region (but these conditions can be changed via function arguments).
Step 5. Remove chimera. The core dada method corrects substitution and indel errors, but chimeras remain. Fortunately, the accuracy of sequence variants after denoising makes identifying chimeric ASVs simpler than when dealing with fuzzy OTUs. Chimeric sequences are identified if they can be exactly reconstructed by combining a left-segment and a right-segment from two more abundant “parent” sequences. The frequency of chimeric sequences varies substantially from dataset to dataset, and depends on on factors including experimental procedures and sample complexity.
1. Read Quality Plots NGS sequence analaysis starts with visualizing the quality of the sequencing. Below are the quality plots of the first sample for the R1 and R2 reads separately. In gray-scale is a heat map of the frequency of each quality score at each base position. The mean quality score at each position is shown by the green line, and the quartiles of the quality score distribution by the orange lines. The forward reads are usually of better quality. It is a common practice to trim the last few nucleotides to avoid less well-controlled errors that can arise there. The trimming affects the downstream steps including error model building, merging and chimera calling. FOMC uses an empirical approach to test many combinations of different trim length in order to achieve best final amplicon sequence variants (ASVs), see the next section “Optimal trim length for ASVs”.
Below is the link to a PDF file for viewing the quality plots for all samples:
2. Optimal trim length for ASVs The final number of merged and chimera-filtered ASVs depends on the quality filtering (hence trimming) in the very beginning of the DADA2 pipeline. In order to achieve highest number of ASVs, an empirical approach was used -
Below is the result of such operation, showing ASV percentages of total reads for all trimming combinations (1st Column = R1 lengths in bases; 1st Row = R2 lengths in bases):