benjjneb
commented
2 years ago From tutorials, the taxonomy was assigned based on st, not st.nochim, is that true? should I proceed similarly in my analysis protocol? I think I should use st.nochim to assign taxonomy and proceed in analyses, correct?
You should probably use st.nochim.
When I checked chimeras from the first project: 535 sequences identified where 182 were bimera (bim true (351), false (184)), and [sum(st[,bim])/sum(st)], I got this (0.06553758). Is it a good value for my sequences?
This is well within expectations. The PCR protocol used in the DADA2+PacBio paper was low cycle number (12 if I remember) and can be expected to have fewer chimeras than higher cycle number protocols. As long as you are using the minFoldParentOverAbundance=3.5 (or higher) than I think you are fine to proceed. PCR produces many chimeras, but mostly at very low frequencies, which is consistent with what you are seeing.
In DADA2 documentation, there is a repeated paragraph in different functions mentioning that (QualityType: (Optional). character(1). The quality encoding of the fastq file(s). "Auto" (the default) means to attempt to auto-detect the encoding. This may fail for PacBio files with uniformly high-quality scores, in which case use "FastqQuality". This parameter is passed on to readFastq; see information there for details. I do not know how to use this parameter and how could I know if I need to use this optional step for my data or not?
If things are running afterwards, you are OK.
Second, regarding Illumina MiSeq sequences (I have previous experiences in other projects), I use Qiime2 for analyses, and I noticed that the percentage input non-chimeric (dada2-stat.qzv) from sequenced samples from the 3 different projects ranged in total between (15-25%) which is too low. However, from a previous project, the majority were above 90%. I am wondering if there is something to do to check the reason behind this huge variability. I checked with the sequencing facility and they responded that everything is okay from their side regarding the sequencing protocol and quality checks. In reality, the quality of the sequences is good when I checked demux_paired_end.qzv. Any explanation, please?
I would need a more complete description of what you are seeing here to try to diagnose, but the cause of the vast majority of issues with too many chimeras being detected is that primers were not removed. Is there a difference in how you are (or aren't) removing primers in these two workflows?
emankhalaf
Hi,
I am new to PacBio technology and I am currently working on 16S sequences generated from both Illumina MiSeq and PacBio Sequel II tech. I have 3 different projects including different plant tissues. I have a few questions regarding the initial analyses of 16S reads.
First, regarding PacBio sequences, I am using DADA2 tutorials provided in (Callahan, et al., 2019: High-throughput amplicon sequencing of the full-length 16S rRNA gene with single-nucleotide resolution). I noticed that in the script of DADA2 + PacBio: Fecal Samples, there were chimeras (39 chimeric sequences), however, the downstream analyses were performed on st.rds, not st.nochim. is that okay?
Second, regarding Illumina MiSeq sequences (I have previous experiences in other projects), I use Qiime2 for analyses, and I noticed that the percentage input non-chimeric (dada2-stat.qzv) from sequenced samples from the 3 different projects ranged in total between (15-25%) which is too low. However, from a previous project, the majority were above 90%. I am wondering if there is something to do to check the reason behind this huge variability. I checked with the sequencing facility and they responded that everything is okay from their side regarding the sequencing protocol and quality checks. In reality, the quality of the sequences is good when I checked demux_paired_end.qzv. Any explanation, please?
Thanks for any help you can provide! Eman