Haitao Wang

University of Greifswald · Institute for Microbiology

We recently got our sequencing data from a company sequenced with Novaseq. As far as I have learnt, Novaseq uses a simpler quanlity score system, which merges different scores into only four categories/levels. This of course causes problem when processing the data with quality-score based pipeline, e.g. DADA2. I got really weired results with DADA2. We also tried with UNOISE3, the ASV table looks fine for the mock data. But since this is also based on quanlity, I'm not sure if this can be trusted.

I also searched for some publications using Novaseq for amplicon sequencing. They all used traditional/old pipelines based on similarity to generate OTUs. I wonder if this is the only way to do this.

I wish there could be someone who has experience dealing with this situation. Is there still some possible way to use DADA2 or UNOISE3 with some modifications to process the data?

I have a dataset, assoicated with 5 time points. For each time point, I have three replicates. I understand that for a time-series dataset, repeated-measures ANOVA or Friedman test would be the best solution.

However, the three replicates are totally random. Let's say, I measured the gas flux from three random locations in a small region (20*20 square meters) at each time point. Even though the measurement was done in the same region, I can't relate these random replicates as three sets.

What I did was using independent ANOVA and using time points as the groups. But I know this is inappropriate. Anyone knows if there is an appropriate method for dealing with this situation?

I have hundreds of samples with different sample sizes/counts (3,000 to 150,000). Due to the uneven sizes, comparing the richness between samples can be tricky without rarefying. However, if I rarefy the data to 3,000 counts per sample, many samples' richness would drop significantly, which is also tricky. I wonder if I should use the rarefied richness or the normal richness for data interpretation.

I personally don't like rarefying, which might throw away many rare taxa. I know Shannon which takes into account the relative abundance of each species can to some extent solve this problem. But what if I'm specifically interested in the richness?

I hope someone could provide some enlightening comments on this.

I have created co-occurrence network based on Spearman's correlations for several times with different projects. My research focuses on microbiome. So the networks were the correlations between different microbial taxa. When I filtered the correlations by coefficients above 0.7 or 0.8, I only got positive correlations (or at least most of them positive correlations). My colleagues also got the same issue as me. After I checked many literatures, they also got like mostly the positive correlations based on filtering with a high cutoff of coefficients.

Is there any reason why this happened? Then how to interpret from all positive correlations?

I'm quite confused about using DESeq2 to find the differential abundant taxa in microbiome studies, especially when there are more than two groups of the factor. I know DESeq2 was initially used for RNA-seq to detect the regulation of gene expressions. It's easy to understand when there are only two groups, e.g. treated vs. untreated. We can easily say which taxa was up-regulated by looking at the log2fold change (positive or negative).

BUT what about when there are three groups, control vs. treat1 vs. treat2? The DESeq2 can still handle this situation, but then I have no idea how to interpret the log2fold change. If we detected some taxa that were significantly different, how can we know in which group these taxa were up-regulated?

Although some people suggest to do the pairwise comparison, it's still unclear to me how to do it and how to interpret it. Does anyone have very good recommendation or idea about this?

Thanks in advance.

I always have some problems with the formula while doing PERMANOVA. For example, adonis(distance.bray ~ x+y, sample.data, permutations = 999) and adonis(distance.bray ~ y+x, sample.data, permutations = 999) gave me the different results of the R square values for factor x and y. So what are the difference between these two formula? And which one should I trust?

Also, I once did the analysis with three factors. When I did adonis(distance.bray ~ x+y+z, sample.data, permutations = 999), z was significant. But when I do adonis(distance.bray ~ z, sample.data, permutations = 999), z was not significant. Can someone explain what happened in these analyses?

I'm wondering if there is a clear method that can be used to create the bipartite network to connect the samples with OTUs/taxa. As I searched the internet, all of the tutorials are talking about how to plot it but not how to calculate or how to get the associations between samples and taxa. Also, the papers that used bipartite network are unclear about the methods either. Although the qiime have the function to get the nodes and edges for bipartite (http://qiime.org/scripts/make_bipartite_network.html), it is unclear how qiime does the analysis. So what is the method/algorithm behind this?

I used adonis to test the difference between groups/categories based on the 16s high-throughput sequencing data. I got a very significant result (p = 0.001), but the R² is very low (< 0.1). I know this means this model only explains less than 10 percent of my data. From the PCoA plot, there were some overlaps between the groups. Do these overlaps contribute to the low R²? I'm wondering if I can still say that the difference between groups is significant? 

I also did ANOSIM and the results are similar. The statistic value is very low (~0.1) but p-value is 0.001. Is this the same problem I got when doing adonis?

Some discussion groups said that there is no relationship between R² and p-value. Doesn't it matter however small the R² is?