RamEx: an R package for high-throughput microbial ramanome analyses with accurate quality assessment.

Area of Science:

• Microbiology
• Spectroscopy
• Bioinformatics

Background:

• Microbial single-cell Raman spectroscopy (SCRS) is vital for label-free microbial phenotyping and community analysis.
• High-throughput SCRS data often suffers from spectral anomalies, hindering accurate classification.
• Development of robust algorithms for outlier detection in microbial ramanome analysis is needed.

Purpose of the Study:

• Introduce RamEx, an R package for high-throughput microbial ramanome analysis.
• Implement the Iterative Convolutional Outlier Detection (ICOD) algorithm for robust spectral anomaly detection.
• Provide a scalable workflow for microbial phenotype differentiation and trait mapping.

Main Methods:

• Developed the Iterative Convolutional Outlier Detection (ICOD) algorithm for dynamic spectral anomaly detection.
• Created the RamEx R package integrating ICOD for quality control and phenotypic classification.
• Benchmarked RamEx on simulated and real microbial datasets, including bacteria and yeast.

Main Results:

• ICOD achieved high F1 scores (0.97 simulated, 0.74 real data), outperforming existing methods.
• RamEx workflow enables phenotype differentiation, taxonomic identification, and metabolic fingerprinting.
• The package processes over one million spectra per hour with C++ acceleration and GPU parallelization.

Conclusions:

• RamEx bridges high-throughput SCRS data and computational analysis for single-cell microbial studies.
• The toolkit facilitates exploration of microbial ecology, metabolism, and antibiotic susceptibility.
• RamEx is available as a free, open-source R package.