Random sampling associated with microbial profiling leads to overestimated stochasticity inference in community...

Area of Science:

• Ecology
• Microbial Ecology
• Community Assembly

Background:

• Understanding community assembly mechanisms is key in ecology.
• Null models help distinguish deterministic from stochastic ecological processes.
• These methods are increasingly applied to microbial communities.

Purpose of the Study:

• To investigate how random sampling issues affect microbial community assembly inference.
• To evaluate the impact of different null community randomization methods on stochasticity estimation.
• To assess the reliability of microbial stochasticity inference, especially for rare subcommunities.

Main Methods:

• Utilized simulated datasets to analyze the effects of random sampling on microbial profiling.
• Implemented various randomization methods for generating null communities.
• Compared the performance of the stochastic ratio and Raup-Crick metrics under sampling biases.

Main Results:

• Random sampling significantly impacts microbial stochasticity inference.
• The effects of sampling bias and randomization methods differ between whole and abundant subcommunities.
• Stochasticity in rare subcommunities is consistently overestimated regardless of randomization method.
• The stochastic ratio metric is more sensitive to sampling issues than the Raup-Crick metric.

Conclusions:

• Caution is advised when inferring microbial stochasticity from beta-diversity, especially for rare taxa.
• The choice of randomization method for null communities is critical.
• Redefining thresholds for deterministic vs. stochastic processes may be necessary.