Microbiome depiction through user-adapted bioinformatic pipelines and parameters

Area of Science:

• Microbiology
• Bioinformatics
• Genomics

Background:

• The microbiome's role in health and disease is increasingly recognized, but variability in bioinformatic protocols hinders clinical application and reproducibility.
• A standardized approach is lacking for processing microbiome genomic data, impacting accurate taxonomic inference.

Purpose of the Study:

• To compare and optimize three common microbiome analysis programs: DADA2, QIIME2, and mothur.
• To develop user-adapted pipelines for processing paired-end amplicon reads to improve compositional inference accuracy and standardize protocols.

Main Methods:

• Two key parameters, maxEE (error threshold) and QTrim (quality score threshold), were isolated across four pipelines.
• The accuracy of sample inference was evaluated using a mock community of known composition.

Main Results:

• Data loss was proportional to parameter stringency and varied by pipeline.
• Increased sequence read retention correlated with higher accuracy in sample inference.
• Optimized parameters improved the approximation of the known mock community composition, reducing false taxa and unaccounted constituents.

Conclusions:

• Microbial composition data from 16S rRNA marker genes require cautious interpretation.
• User-adapted bioinformatic protocols are essential for accurate microbial community profiling.
• Analysis parameters, pipelines, and target amplicons must be carefully considered for reliable results.