I have extracted DNA from suspended biomass collected on filters. Now I want to divide the OTU counts by the volume of water filtered for each sample, in order to abtain semi-quantitative abundance estimates. Does anyone know a reasonalble way to do it? I'm using the softaware packages "mothur" and "phyloseq". Thus far I only found the option to normalize all samples to one value. In my case, however, I have different values (volumes of filtered water) for each sample.
Thanks for any suggestions,
Yuki.
Hi, it will be complex to relate your number of OTUs back to the amount of water filtered, given you probably used different amounts of template for your PCR and then took a subsample of that for the sequencing. I would have just subsampled the number of sequences per sample to the same amount and use relative abundance
Hi Will and Vani,
Thanks for your suggestions. I was finally able to manipulate the OTU table in R, so the technica part is solved. Regarding the normalisation procedure, you're right that I have to take into account all the library preparation steps as well. I pretty much know the fraction of the extracted DNA that we sent out for PCR and sequencing. I would assume that all samples were treated in exactly the same way by the sequencing lab (right?).
The rational behind this is that I want to correlate OTU abundances (per liter of lake water) with concentrations of specific lipids (per liter). I'm afraid that relative abundances are not really compareble beacause my lipid profiling was very selective...
Do read this paper https://www.ncbi.nlm.nih.gov/pubmed/27255738
Absolute abundances from NGS platforms appear to be not the best way to look at the quantity. It is more a function of the sequencing platform and so sum-constrained data.
- Badges
- Science method