Currently I am analyzing amplicon sequencing data generated from diseased and control individuals. In that observed OTUs are more in control individuals but Shannon index is significantly higher in diseased persons. At the same time, Phylogenetic diversity (PD_whole_tree) calculated in qiime is higher in control individuals.
My question is if observed OTUs are more in controls than how can species richness calculated by Shannon is higher in diseased persons? Is it possible?
Thanks a lot in advance
Cheers
The Shannon index is a measure of diversity, not richness. It measures the number of OTUs in your sample (richness) but scales them based on the evenness of the community. If the controls have more OTUs but a small number of those OTUs dominate the sample they will report a lower Shannon diversity than a a community with fewer OTUs evenly distributed. I've attached a small Excel example of this.
There's a formula to calculate the Shannon evenness of each sample, it basically compares your Shannon diversity to the maximum possible for your sampled community. If you apply this to your samples it might reveal that community unevenness is the cause of the low Shannon score.
The Shannon index is a measure of diversity, not richness. It measures the number of OTUs in your sample (richness) but scales them based on the evenness of the community. If the controls have more OTUs but a small number of those OTUs dominate the sample they will report a lower Shannon diversity than a a community with fewer OTUs evenly distributed. I've attached a small Excel example of this.
There's a formula to calculate the Shannon evenness of each sample, it basically compares your Shannon diversity to the maximum possible for your sampled community. If you apply this to your samples it might reveal that community unevenness is the cause of the low Shannon score.
Thank You so much David,
My all doubts are clear now. The sheet you provided was the simple yet best example providing idea about shannon index.
Can you please brief me about faith's phylogenetic diversity?
Have a nice day
Regards,
I'm less clear on the calculations for Faith's PD, but it's something like the sum of all branch lengths of the OTUs in the group. It would make sense that a sample with more unique OTUs would have a higher PD since they will contribute more of the branches on the tree and therefore would be more likely to have a higher score.
You could potentially get a higher PD for a sample with less OTUs if the OTUs were deep branching but it would depend upon the exact structure/samples.
Personally I'm not a fan of PD, because it's extremely sensitive to subsampling. In my experience scores like Shannon/Simpson are more robust. but there are questions those can't answer so I understand the need for PD.
Hi David,
As far as I understand, PD is calculated based on where the bugs are in the evolution tree of life. If the sample have many bugs from many difference branches they would be more diversity. Said if both samples have 3 outs one ( Nitrospirae, E. coli, NC10) and other sample (E.coli. Shigella, Samonella) then the first one should have way higher in alpha diversity than the second one. It we are assuming that sequencing error can promote quite amount of fake otus (which definitely could happen), the chance of these fake otu to be genetically similar (very close in the tree of life) should be much higher than the chance of fake otus to be far from each other in the tree of life (it would need more mutations). For that I think PD whole tree for alpha-diversity and weighted unifac for beta- diversity would be more robust than other methods. Please correct me if I am wrong.
- Badges
- Science topic
- Similar topics
- Biological Science
- Evolutionary Biology
- Phylogenetics