Computation Subtraction-based Microbiome Discovery (CSMD)

Computation Subtraction-based Microbiome Discovery (CSMD) performs computational subtraction on whole genome shotgun sequencing data to generate species-level microbiome profiles from clinical samples with low microbial biomass by separating microbial DNA from human host DNA and contaminating DNA for accurate detection.

Key Features:

• Maximal Host Sequence Elimination: Maximizes removal of host sequences while preserving microbial signals to enable detection of low-abundance microorganisms.
• High Sensitivity and Specificity: Uses a stepwise convergent solution to increase sensitivity in removing host DNA and specificity in identifying microbial species, yielding more precise abundance estimates.
• Broad Microbial Diversity Detection: Recovers a wider range of microorganisms, including from cell-free plasma DNA, revealing microbial diversity beyond previous characterizations despite contaminating DNA and large RefSeq databases.

Scientific Applications:

• Disease Diagnosis and Prognosis: Provides detailed species-level microbiome profiles that can be used to identify microbial signatures associated with specific diseases or clinical conditions.
• Microbial Ecology Studies: Enables characterization of diversity and dynamics of microbial communities within clinical samples, informing host–microbe interaction studies.
• Personalized Medicine: Supplies accurate microbiome composition data that can inform individualized treatment strategies based on patient-specific microbial profiles.

Methodology:

Operates on whole genome shotgun sequencing data using a series of libraries that include human-related genomes and all RefSeq bacterial genomes; processes genomic datasets stored in directories such as DBPATH/hg38/SEQ and DBPATH/AHG/SEQ, saves indexed files accordingly, and has reported computational requirements of approximately 500 GB additional disk space and 200 GB RAM.