Project FOMC24059 services include NGS sequencing of the V1V9 region of the 16S rRNA gene amplicons from the samples. First and foremost, please download this report, as well as the sequence raw data from the download links provided below. These links will expire after 60 days. We cannot guarantee the availability of your data after 60 days.

Full Bioinformatics analysis service was requested. We provide many analyses, starting from the raw sequence quality and noise filtering, pair reads merging, as well as chimera filtering for the sequences, using the DADA2 denosing algorithm and pipeline.

We also provide many downstream analyses such as taxonomy assignment, alpha and beta diversity analyses, and differential abundance analysis.

For taxonomy assignment, most informative would be the taxonomy barplots. We provide an interactive barplots to show the relative abundance of microbes at different taxonomy levels (from Phylum to species) that you can choose.

If you specify which groups of samples you want to compare for differential abundance, we provide both ANCOM and LEfSe differential abundance analysis.

The samples were processed and analyzed with the ZymoBIOMICS® Service: Targeted Metagenomic Sequencing (Zymo Research, Irvine, CA).

DNA Extraction: If DNA extraction was performed, the following DNA extraction kit was used according to the manufacturer’s instructions:

Targeted Library Preparation: The DNA samples were prepared for targeted sequencing with the Quick-16S™ NGS Library Prep Kit (Zymo Research, Irvine, CA). These primers were custom designed by Zymo Research to provide the best coverage of the 16S gene while maintaining high sensitivity. The primer sets used in this project are marked below:

The samples were processed and analyzed with the ZymoBIOMICS® Targeted Sequencing Service (Zymo Research, Irvine, CA).

DNA Extraction: One of two different DNA extraction kits was used depending on the sample type and sample volume. In most cases, the ZymoBIOMICS®-96 MagBead DNA Kit (Zymo Research, Irvine, CA) was used to extract DNA using an automated platform. In some cases, ZymoBIOMICS® DNA Miniprep Kit (Zymo Research, Irvine, CA) was used for low biomass samples, such as skin swabs, as it permits for a lower elution volume, resulting in more concentrated DNA samples.

Targeted Library Preparation: The library of whole 16S sequencing was prepared by following the full-length 16S amplification protocol from PacBio ( https://www.pacb.com/wp-content/uploads/Procedure-checklist-Amplification-of-bacterial-full-length-16S-rRNA-gene-with-barcoded-primers.pdf). In brief, the whole 16S gene was amplified using the 27f (AGRGTTYGATYMTGGCTCAG) and 1492r(RGYTACCTTGTTACGACTT) primers with barcodes and adapters. 2ng DNA was used as the PCR template for each sample. The PCR was run with 25 cycles following the conditions mentioned in the protocol. After that, the PCR product of each reaction underwent clean up with Select-a-Size DNA Clean & Concentrator MagBead Kit (Zymo Research, Irvine, CA) keeping fragments >300bp. The library of each reaction is then quantified using NanoDrop and pooled together with equal DNA mass. The pooled library was prepared for PacBio sequencing using the SMRTbell® prep kit 3.0.

Control Samples: The ZymoBIOMICS® Microbial Community Standard (Zymo Research, Irvine, CA) was used as a positive control for each batch of DNA extraction. The ZymoBIOMICS® Microbial Community DNA Standard (Zymo Research, Irvine, CA) was also used as a positive control for the library preparation process. Negative controls (i.e. blank extraction control, blank library preparation control) were included to assess the level of bioburden carried by the wet-lab process.

Sequencing: The library was sequenced on one 8M SMRT cell on the Sequel IIe system.

The sequencing library was prepared using an innovative library preparation process in which PCR reactions were performed in real-time PCR machines to control cycles and therefore limit PCR chimera formation. The final PCR products were quantified with qPCR fluorescence readings and pooled together based on equal molarity. The final pooled library was cleaned up with the Select-a-Size DNA Clean & Concentrator™ (Zymo Research, Irvine, CA), then quantified with TapeStation® (Agilent Technologies, Santa Clara, CA) and Qubit® (Thermo Fisher Scientific, Waltham, WA).

Control Samples: The ZymoBIOMICS® Microbial Community Standard (Zymo Research, Irvine, CA) was used as a positive control for each DNA extraction, if performed. The ZymoBIOMICS® Microbial Community DNA Standard (Zymo Research, Irvine, CA) was used as a positive control for each targeted library preparation. Negative controls (i.e. blank extraction control, blank library preparation control) were included to assess the level of bioburden carried by the wet-lab process.

Sequencing: The final library was sequenced on Illumina® NextSeq 2000™ with a p1 (Illumina, Sand Diego, CA) reagent kit (600 cycles). The sequencing was performed with 25% PhiX spike-in.

Bioinformatics Analysis of Sequence Data: Raw full-length 16S V1V9 PacBio amplicon sequences obtained from Zymo Research were subject to primer removal, quality-filtered, denoised, pair-end merged, and chimera removed with the DADA2 tool (version 1.12.1) [1]. Amplicon sequence variants (ASVs) generated by DADA2 were subjected to species level taxonomy assignment based on the approach developed by Al-Hebshi et. al. [2] against the FOMC Reference Sequence Set (https://microbiome.forsyth.org/ftp/refseq) that consists of 1,015 full-length 16S rRNA sequences from HOMD V15.22, 356 from MOMD V5.1, and 22,126 from NCBI, a total of 23,497 sequences. Species-level taxonomy assignment algorithm is available at https://github.com/tsute/FOMC_16S_rRNA_Taxonomy_Assignment_Algorithm. Altogether these sequences represent a total of 17,035 oral and non-oral microbial species. Specie-level read count tables were imported using R’s “phyloseq” package (version 1.28.0) [3] for downstream analyses. Alpha diversity was calculated in 3 measurements: 1) number of species (observed), 2) Shannon index, and 3) Simpson index using “phyloseq” package’s “plot_richness” function [3]. Alpha diversity significance tests were evaluated with QIIME2 (version 2020.11) “alpha-group-significance” function [4]. Beta diversity NMDS plots were generated with the “ordinate” function in “phyloseq” and beta diversity significance tests were evaluated with QIIME2 “beta-group-significance” function [4]. ANCOMB-BC2 (Analysis of Compositions of Microbiomes with Bias Correction) (version 2.0.2) was used to test the significance of differential abundance between two test groups [5]. Linear discriminant analysis LDA Effect Size (LEfSe) (version 1.0.0) was used to plot the effect size of differentially abundant features [6]. The FOMC analysis pipepline software version information is available at https://microbiome.forsyth.org/software.php. References (for Bioinformatics Analysis of Sequence Data):

1. Callahan BJ, McMurdie PJ, Rosen MJ, Han AW, Johnson AJ, Holmes SP. DADA2: High-resolution sample inference from Illumina amplicon data. Nat Methods. 2016 Jul;13(7):581-3. doi: 10.1038/nmeth.3869. Epub 2016 May 23. PMID: 27214047; PMCID: PMC4927377.
2. Al-Hebshi NN, Nasher AT, Idris AM, Chen T. Robust species taxonomy assignment algorithm for 16S rRNA NGS reads: application to oral carcinoma samples. J Oral Microbiol. 2015 Sep 29;7:28934. doi: 10.3402/jom.v7.28934. PMID: 26426306; PMCID: PMC4590409.
3. McMurdie PJ, Holmes S. phyloseq: an R package for reproducible interactive analysis and graphics of microbiome census data. PLoS One. 2013 Apr 22;8(4):e61217. doi: 10.1371/journal.pone.0061217. PMID: 23630581; PMCID: PMC3632530.
4. Bolyen et al. Reproducible, interactive, scalable and extensible microbiome data science using QIIME 2. Nat Biotechnol. 2019 Aug;37(8):852-857. doi: 10.1038/s41587-019-0209-9. Erratum in: Nat Biotechnol. 2019 Sep;37(9):1091. doi: 10.1038/s41587-019-0252-6. PMID: 31341288; PMCID: PMC7015180.
5. Mandal S, Van Treuren W, White RA, Eggesbø M, Knight R, Peddada SD. Analysis of composition of microbiomes: a novel method for studying microbial composition. Microb Ecol Health Dis. 2015 May 29;26:27663. doi: 10.3402/mehd.v26.27663. PMID: 26028277; PMCID: PMC4450248.
6. Segata N, Izard J, Waldron L, Gevers D, Miropolsky L, Garrett WS, Huttenhower C. Metagenomic biomarker discovery and explanation. Genome Biol. 2011 Jun 24;12(6):R60. doi: 10.1186/gb-2011-12-6-r60. PMID: 21702898; PMCID: PMC3218848.

The complete report of your project, including all links in this report, can be downloaded by clicking the link provided below. The downloaded file is a compressed ZIP file and once unzipped, open the file “REPORT.html” (may only shown as "REPORT" in your computer) by double clicking it. Your default web browser will open it and you will see the exact content of this report.

Please download and save the file to your computer storage device. The download link will expire after 60 days upon your receiving of this report.

Complete report download link:

To view the report, please follow the following steps:

The raw NGS sequence data is available for download with the link provided below. The data is a compressed ZIP file and can be unzipped to individual sequence files. Since this is a pair-end sequencing, each of your samples is represented by two sequence files, one for READ 1, with the file extension “*_R1.fastq.gz”, another READ 2, with the file extension “*_R1.fastq.gz”. The files are in FASTQ format and are compressed. FASTQ format is a text-based data format for storing both a biological sequence and its corresponding quality scores. Most sequence analysis software will be able to open them. The Sample IDs associated with the R1 and R2 fastq files are listed in the table below:

Sample ID
Original Sample ID
Read File Name

F24059.S10
original sample ID here
F24059.S10.fastq.gz

F24059.S11
original sample ID here
F24059.S11.fastq.gz

F24059.S12
original sample ID here
F24059.S12.fastq.gz

F24059.S13
original sample ID here
F24059.S13.fastq.gz

F24059.S14
original sample ID here
F24059.S14.fastq.gz

F24059.S15
original sample ID here
F24059.S15.fastq.gz

F24059.S16
original sample ID here
F24059.S16.fastq.gz

F24059.S17
original sample ID here
F24059.S17.fastq.gz

F24059.S18
original sample ID here
F24059.S18.fastq.gz

F24059.S19
original sample ID here
F24059.S19.fastq.gz

F24059.S01
original sample ID here
F24059.S01.fastq.gz

F24059.S20
original sample ID here
F24059.S20.fastq.gz

F24059.S21
original sample ID here
F24059.S21.fastq.gz

F24059.S22
original sample ID here
F24059.S22.fastq.gz

F24059.S23
original sample ID here
F24059.S23.fastq.gz

F24059.S02
original sample ID here
F24059.S02.fastq.gz

F24059.S03
original sample ID here
F24059.S03.fastq.gz

F24059.S04
original sample ID here
F24059.S04.fastq.gz

F24059.S05
original sample ID here
F24059.S05.fastq.gz

F24059.S06
original sample ID here
F24059.S06.fastq.gz

F24059.S07
original sample ID here
F24059.S07.fastq.gz

F24059.S08
original sample ID here
F24059.S08.fastq.gz

F24059.S09
original sample ID here
F24059.S09.fastq.gz

Please download and save the file to your computer storage device. The download link will expire after 60 days upon your receiving of this report.

Raw sequence data download link:

What is DADA2?

DADA2 is a software package that models and corrects Illumina-sequenced amplicon errors [1]. DADA2 can also be used to correct errors in PacBio CCS amplicon sequencing data, which was used in this project for the V1V9 full-length 16S rRNA sequencing. 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

References

1. Callahan BJ, McMurdie PJ, Rosen MJ, Han AW, Johnson AJ, Holmes SP. DADA2: High-resolution sample inference from Illumina amplicon data. Nat Methods. 2016 Jul;13(7):581-3. doi: 10.1038/nmeth.3869. Epub 2016 May 23. PMID: 27214047; PMCID: PMC4927377.

Analysis Procedures:

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. Primer removal and sequence quality filtering Primers were removed from the sequences using the "removePrimers" function in the DADA2 package, to remove the primer sequences F27 (AGRGTTYGATYMTGGCTCAG) and R1492 (RGYTACCTTGTTACGACTT) from both ends of the reads. DADA2 allows to trim off the poor quality reads in order to improve the error model building and pair merging performance using the "filterAndTrim" function in DADA2 package with these parameters: minQ=3, minLen=1000, maxLen=1600, maxN=0, rm.phix=FALSE, maxEE=2.

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). Prior to the error learning, the read pairs were dereplicated first using the "derepFastq" function in the DADA2 package.

Step 3. Infer amplicon sequence variants (ASVs) based on the error model built in previous step. This step is also called sequence "denoising". The sequences were subject to the "dada" denoising function using the error model learned from previous step. The outcome of this step is a list of ASVs that are the equivalent of oligonucleotides.

Step 4. 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. The chimera were removed using the "removeBimeraDenovo" function in the DADA2 package.

Results

1. Read Quality Plots NGS sequence analysis 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”.

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 -

1. Create a random subset of each sample consisting of 5,000 R1 and 5,000 R2 (to reduce computation time)
2. Trim 10 bases at a time from the ends of both R1 and R2 up to 50 bases
3. For each combination of trimmed length (e.g., 300x300, 300x290, 290x290 etc), the trimmed reads are subject to the entire DADA2 pipeline for chimera-filtered merged ASVs
4. The combination with highest percentage of the input reads becoming final ASVs is selected for the complete set of data

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):

3. Error plots from learning the error rates After DADA2 building the error model for the set of data, it is always worthwhile, as a sanity check if nothing else, to visualize the estimated error rates. The error rates for each possible transition (A→C, A→G, …) are shown below. Points are the observed error rates for each consensus quality score. The black line shows the estimated error rates after convergence of the machine-learning algorithm. The red line shows the error rates expected under the nominal definition of the Q-score. The ideal result would be the estimated error rates (black line) are a good fit to the observed rates (points), and the error rates drop with increased quality as expected.

Forward Read R1 Error Plot


Reverse Read R2 Error Plot

The PDF version of these plots are available here:

4. DADA2 Result Summary The table below shows the summary of the DADA2 analysis, tracking paired read counts of each samples for all the steps during DADA2 denoising process - including end-trimming (filtered), denoising (denoisedF, denoisedF), pair merging (merged) and chimera removal (nonchim).

This table can be downloaded as an Excel table below:

5. DADA2 Amplicon Sequence Variants (ASVs). A total of 1928 unique merged and chimera-free ASV sequences were identified, and their corresponding read counts for each sample are available in the "ASV Read Count Table" with rows for the ASV sequences and columns for sample. This read count table can be used for microbial profile comparison among different samples and the sequences provided in the table can be used to taxonomy assignment.

The table can be downloaded from this link:

The close-reference taxonomy assignment of the ASV sequences using BLASTN is based on the algorithm published by Al-Hebshi et. al. (2015)[2].

The species-level, open-reference 16S rRNA NGS reads taxonomy assignment pipeline Version 20210310a

1. Raw sequences reads in FASTA format were BLASTN-searched against a combined set of 16S rRNA reference sequences - the FOMC 16S rRNA Reference Sequences version 20221029 (https://microbiome.forsyth.org/ftp/refseq/). This set consists of the HOMD (version 15.22 http://www.homd.org/index.php?name=seqDownload&file&type=R ), Mouse Oral Microbiome Database (MOMD version 5.1 https://momd.org/ftp/16S_rRNA_refseq/MOMD_16S_rRNA_RefSeq/V5.1/), and the NCBI 16S rRNA reference sequence set (https://ftp.ncbi.nlm.nih.gov/blast/db/16S_ribosomal_RNA.tar.gz). These sequences were screened and combined to remove short sequences (<1000nt), chimera, duplicated and sub-sequences, as well as sequences with poor taxonomy annotation (e.g., without species information). This process resulted in 1,015 full-length 16S rRNA sequences from HOMD V15.22, 356 from MOMD V5.1, and 22,126 from NCBI, a total of 23,497 sequences. Altogether these sequence represent a total of 17,035 oral and non-oral microbial species. The NCBI BLASTN version 2.7.1+ (Zhang et al, 2000) [3] was used with the default parameters. Reads with ≥ 98% sequence identity to the matched reference and ≥ 90% alignment length (i.e., ≥ 90% of the read length that was aligned to the reference and was used to calculate the sequence percent identity) were classified based on the taxonomy of the reference sequence with highest sequence identity. If a read matched with reference sequences representing more than one species with equal percent identity and alignment length, it was subject to chimera checking with USEARCH program version v8.1.1861 (Edgar 2010). Non-chimeric reads with multi-species best hits were considered valid and were assigned with a unique species notation (e.g., spp) denoting unresolvable multiple species. 2. Unassigned reads (i.e., reads with < 98% identity or < 90% alignment length) were pooled together and reads < 200 bases were removed. The remaining reads were subject to the de novo operational taxonomy unit (OTU) calling and chimera checking using the USEARCH program version v8.1.1861 (Edgar 2010)[4]. The de novo OTU calling and chimera checking was done using 98% as the sequence identity cutoff, i.e., the species-level OTU. The output of this step produced species-level de novo clustered OTUs with 98% identity. Representative reads from each of the OTUs/species were then BLASTN-searched against the same reference sequence set again to determine the closest species for these potential novel species. These potential novel species were pooled together with the reads that were signed to specie-level in the previous step, for down-stream analyses. The taxonomy assignment algorithm is available at github https://github.com/tsute/FOMC_16S_rRNA_Taxonomy_Assignment_Algorithm Reference: Al-Hebshi NN, Nasher AT, Idris AM, Chen T. Robust species taxonomy assignment algorithm for 16S rRNA NGS reads: application to oral carcinoma samples. J Oral Microbiol. 2015 Sep 29;7:28934. doi: 10.3402/jom.v7.28934. PMID: 26426306; PMCID: PMC4590409. Zhang Z, Schwartz S, Wagner L, Miller W. A greedy algorithm for aligning DNA sequences. J Comput Biol. 2000 Feb-Apr;7(1-2):203-14. doi: 10.1089/10665270050081478. PMID: 10890397. Edgar RC. Search and clustering orders of magnitude faster than BLAST. Bioinformatics. 2010 Oct 1;26(19):2460-1. doi: 10.1093/bioinformatics/btq461. Epub 2010 Aug 12. PubMed PMID: 20709691. 3. Designations used in the taxonomy: 1) Taxonomy levels are indicated by these prefixes: k__: domain/kingdom p__: phylum c__: class o__: order f__: family g__: genus s__: species Example: k__Bacteria;p__Firmicutes;c__Clostridia;o__Clostridiales;f__Lachnospiraceae;g__Blautia;s__faecis 2) Unique level identified – known species: k__Bacteria;p__Firmicutes;c__Clostridia;o__Clostridiales;f__Lachnospiraceae;g__Roseburia;s__hominis The above example shows some reads match to a single species (all levels are unique) 3) Non-unique level identified – known species: k__Bacteria;p__Firmicutes;c__Clostridia;o__Clostridiales;f__Lachnospiraceae;g__Roseburia;s__multispecies_spp123_3 The above example “s__multispecies_spp123_3” indicates certain reads equally match to 3 species of the genus Roseburia; the “spp123” is a temporally assigned species ID. k__Bacteria;p__Firmicutes;c__Clostridia;o__Clostridiales;f__Lachnospiraceae;g__multigenus;s__multispecies_spp234_5 The above example indicates certain reads match equally to 5 different species, which belong to multiple genera.; the “spp234” is a temporally assigned species ID. 4) Unique level identified – unknown species, potential novel species: k__Bacteria;p__Firmicutes;c__Clostridia;o__Clostridiales;f__Lachnospiraceae;g__Roseburia;s__ hominis_nov_97% The above example indicates that some reads have no match to any of the reference sequences with sequence identity ≥ 98% and percent coverage (alignment length) ≥ 98% as well. However this groups of reads (actually the representative read from a de novo OTU) has 96% percent identity to Roseburia hominis, thus this is a potential novel species, closest to Roseburia hominis. (But they are not the same species). 5) Multiple level identified – unknown species, potential novel species: k__Bacteria;p__Firmicutes;c__Clostridia;o__Clostridiales;f__Lachnospiraceae;g__Roseburia;s__ multispecies_sppn123_3_nov_96% The above example indicates that some reads have no match to any of the reference sequences with sequence identity ≥ 98% and percent coverage (alignment length) ≥ 98% as well. However this groups of reads (actually the representative read from a de novo OTU) has 96% percent identity equally to 3 species in Roseburia. Thus this is no single closest species, instead this group of reads match equally to multiple species at 96%. Since they have passed chimera check so they represent a novel species. “sppn123” is a temporary ID for this potential novel species.

4. The taxonomy assignment algorithm is illustrated in this flow char below: