Thresher: an improved algorithm for peak height thresholding of microbial community profiles

Area of Science:

• Microbiology
• Bioinformatics
• Molecular Ecology

Background:

• Automated rRNA intergenic spacer analysis (ARISA) is widely used for microbial community profiling.
• Conventional ARISA methods often apply a single threshold across all samples, which can compromise accuracy due to variations in community richness.
• This limitation hinders reliable cross-sample comparisons and accurate data interpretation.

Purpose of the Study:

• To introduce Thresher, an improved technique for determining peak height thresholds in ARISA profiles.
• To address the limitations of sample-independent thresholding in fragment analysis.
• To provide a robust method for outlier rejection in ARISA data.

Main Methods:

• Thresher calculates sample-dependent thresholds individually for each replicate within a pair and for each sample.
• Thresholds are selected to minimize dissimilarity between replicates post-thresholding.
• A quantitative similarity test is employed to identify and reject invalid replicates or samples.

Main Results:

• The Thresher algorithm demonstrates superior performance compared to conventional thresholding techniques.
• Thresholding is performed individually for each sample, accounting for community richness variations.
• The method effectively identifies and rejects non-valid replicate pairs, improving data quality.

Conclusions:

• Thresher offers a more accurate and reliable approach to ARISA data analysis.
• Sample-dependent thresholding is crucial for robust microbial community profiling.
• The developed algorithm enhances the comparability and integrity of ARISA-based ecological studies.