unknown ASVs and dust dispersal

While taxonomically unassigned ASVs can be disregarded as useless, they can in fact still hold interesting ecological information. In a paper from my old lab, my former colleague Yongjian, while studying patchy desert vegetation, found that bare desert soils hold a higher proportion of unannotated 16S rRNA phylotypes (or ASVs) compared to vegetated soils. He argued that bacteria living in bare soils are oligotrophs (they do not need much resources but grow slower) and are harder to cultivate, while copiotrophs (fast grower, resource inefficient bacteria), which thrive in vegetated soil, are easier to cultivate, as they tendentiously grow faster and adapt to a wider ranges of sources than oligotrophs. This reflects into a higher proportion of copiotrophs to be easily cultivated, a higher availability of references genomes, and a better taxonomic coverage.

Similar trend happens also in our sky-islands study. In this paper, we sampled soil microbial communities along eight different elevation gradients in the Madrean sky islands. These are mountains were you have forest at the top and desert at the bottom, due to the unique Arizonan climatic history.

If we look at a figure here below, I plotted the relative summed relative abundance of all ASVs with unknown annotations against the elevation of the sample. We see that desert soils (lower elevation) have higher abundance of unannotated ASVs compared to high elevation soils, characterized by pine and broadleaves forests. Although this trend is only visible at low taxonomic resolutions.

The results suggests that similar ecological dynamics happen at also bigger (landscape) scales.

So I went on and look at my dust data. In my dust study, I collected bare soil from four points in the desert, and using a wind-tunnel (WT) I collected also dust that is potentially generated from this points. I then looked at the differences in their composition when compared to soil samples. If I sum up the relative abundance of unknown, I get opposite trends than expected.

What I found is a higher proportion of unassigned ASVs in soil compared to wind tunnel (WT) dust. This seem to contradict the previous two results as I expected dust to be more oligotrophic than soil, but it seems in fact to be the other way around. These result seem to suggest that dust particles are actually populated by a more copiotrophic community in comparison to the sampled surface soil (which was just about the first cm of soil). So this further reinforce the story of the paper, that dust dispersed bacteria are well adapted for dust dispersal as a part of their life strategy, as usually generalists/copiotrophs are thought to have higher chances to be good dispersers.

Further on, when I will have time to work again on this, I would like to check which taxonomic groups hold a higher proportion of unassigned...I know it sounds like a contradiction...but stay tuned.

You can find the code here.