I am wondering what would be the best way to compare a pair of samples. I would be interested to study whether one sample (a child’s sample) is more like its pair (the mother’s sample) than other (children’s) samples. I read this article (https://docs.qiime2.org/2018.11/tutorials/longitudinal/?highlight=long#pairwise-distance-comparisons) but was unsure what to do so I thought I would ask first. Thank you in advance for the tips.
Are you looking at a single sample (n=1) or are there replicates for each case? If you are just looking at a single sample you don’t really need (read: can’t do) statistical testing but you could use a variety of analyses to compare. For example, Alpha diversity measures can tell you how rich/even your one sample is compared to the other samples but says nothing about the actual composition of the community. S simple PCoA plot can show dis(similarity) between your samples that take into consideration composition and is pretty easy to digest given its data-reduction approach. Dots closer to each other are more similar to each other. Creating a bar graph or heat maps can also give a nice visual of comparing the different individual samples. The longitudinal plugin you cited, much like most statistical tests, requires multiple samples in a category for it to calculate means and variance and compare those across time/groups.
Thank you for the reply! I have multiple samples from the same child and one from the mother, and I’m wondering if it would be possible to compare the child’s samples to the mother’s. PCoA sounds like a good approach. Is there a tutorial for creating a heat map?