I recently started working with 16s metabarcoding and I just encountered a situation I don't know it's common or not. I'm mostly working with the bioinformatics part of the project.
We have 5 sampling times (different years) and after processing the fastq files with dada2 (same parameters) and using phyloseq (I decided to use rarefy_even_depth despite the reservations from the phyloseq authors), we found out that observed richness varied significantly with the year. The first year, treatments had 1000 ASVs; the following three years had 2500-3000 ASVs; the last year, it was back to around 1000 ASVs.
Other factors involved:
a) the five sampling times were sequenced in different runs and different machines (same model, but different machines. That doesn't explain why three were in the same ball park, while another two were similar as well.
b) the fastq files concerning the last sampling time were provided with primers, so I used cutadapt to remove them. That doesn't explain why the first sampling time (provided without primers) and the last sampling time were similar.
I wonder if those results are legit and variation is due to edafoclimatic conditions and crop cultivated during sampling time.
P.s.: As a side note, when plotting a PCoA with bray-curtis dissimilarity, all sampling times are clearly differentiated.
It could be due to a combination of factors, including edafoclimatic conditions, crop cultivated during sampling time, and technical variations in sequencing and data processing. Few points to consider:
1. Biological Factors: Edafoclimatic conditions and the crop cultivated during sampling time can indeed influence the microbial community composition and diversity. Different years may exhibit variations in environmental factors such as temperature, precipitation, soil composition, and plant-associated microbiota, which can impact microbial diversity.
2. Technical Variations: The use of different sequencing runs and machines may introduce technical variations in the data. Although the machines are of the same model, subtle differences in protocols, reagents, and sequencing chemistry can affect the sequencing outcomes. It's important to note that such technical variations can contribute to the observed differences in richness across sampling years.
3. Primers and Cutadapt: The presence of primers in the fastq files of the last sampling time and their subsequent removal with Cutadapt could potentially introduce biases or affect the downstream analysis. Primer sequences can influence PCR amplification efficiency and specificity, which may impact the observed richness. However, it's worth noting that the similarity between the first and last sampling times suggests that primer presence alone may not explain the observed differences in richness.
4. Data Processing: The use of the rarefy_even_depth function in Phyloseq can further influence the observed richness. Rarefaction is a technique used to standardize sequencing depth across samples, but it can reduce the number of observed ASVs and potentially impact diversity estimates. The decision to rarefy_even_depth should be made carefully, considering the specific goals of your analysis and potential caveats associated with rarefaction.
5. PCoA and Bray-Curtis Dissimilarity: The differentiation of sampling times in the PCoA plot based on Bray-Curtis dissimilarity supports the notion that the microbial communities exhibit temporal dynamics. This further suggests that the observed variations in richness could reflect genuine differences in community composition across sampling years.
Hope it helps
- Badges
- Science topic
- Similar topics
- Physical Sciences
- Materials Science
- Plating