Improving an rRNA depletion protocol with statistical design of experiments

In prokaryotic RNA-seq library preparation, rRNA depletion is required to remove highly abundant rRNA transcripts from total RNA. rRNA is so abundant that small improvements in depletion efficiency lead to large increases in mRNA sequencing coverage. The current gold-standard method for rRNA depletion makes rRNA depletion the most expensive step in prokaryotic RNA-seq library preparation. A variety of commercial and home-made methods exist to lower the cost or increase the efficiency of rRNA removal. Many of these techniques are suboptimal when applied to new species of bacteria or when the protocol or reagents need to be changed. Re-optimizing a protocol by trial-and-error is an expensive and laborious process. Systematic frameworks like the statistical design of experiments (DOE) can efficiently improve processes by exploring the quantitative relationship between multiple factors. DOE allows experimenters to find factor interactions that may not be apparent when factors are studied in isolation.

Researchers at the University of Illinois Urbana-Champaign used DOE to optimize an rRNA depletion protocol by updating reagents and identifying factors that maximize rRNA removal and minimize cost. The optimized protocol more efficiently removes rRNA, uses fewer reagents, and is less expensive than the original protocol. This optimization required only 36 experiments and identified two significant interactions among three protocol factors. Overall, this approach demonstrates the utility of a rational, DOE framework for improving complex molecular biology protocols.

Response surface methodology (RSM) is used for process optimization

A. RSM is performed in three stages. First, a multi-factorial experimental design is selected. Second, the experiment is carried out, the response is measured, and a mathematical model is fit to the response surface. Third, the model guides the search for optimal settings. B. A 3-factor rotatable central composite design (CCD) was chosen to assess first-order, two-way interaction, and quadratic effects. The CCD has a factorial core to map first-order and two-way interaction effects, and center and axial points to measure quadratic effects. For a rotatable CCD, the coded level for the axial points (α) is set at F4, where F is the number of factorial points.

David BM, Jensen PA. (2022 )Improving an rRNA depletion protocol with statistical design of experiments. SLAS Technology [Epub ahead of print]. [article]

Leave a Reply

Your email address will not be published. Required fields are marked *


Time limit is exhausted. Please reload CAPTCHA.