A relative-entropy algorithm for genomic fingerprinting captures host-phage similarities.

Area of Science:

• Bioinformatics
• Genomics
• Molecular Biology

Background:

• Codon degeneracy allows multiple DNA sequences to encode the same protein.
• Organism-specific DNA sequences contain hidden, biologically significant short oligonucleotides.
• Previous methods lacked efficient ways to identify these novel genomic signals.

Purpose of the Study:

• To develop and apply an information-theoretic algorithm for discovering novel, biologically significant oligonucleotides.
• To identify unique genomic "fingerprints" for bacterial species.
• To establish a new basis for phylogenetic analysis and classification of environmental DNA fragments.

Main Methods:

• Developed an information-theoretic algorithm to detect short oligonucleotides (2-7 nucleotides).
• Applied the algorithm to 209 bacterial genomes from the NCBI database.
• Introduced a scoring algorithm for accurate species identification and host-phage relationship determination.

Main Results:

• Identified over 100 previously undiscovered, biologically significant oligonucleotides per bacterium.
• Developed species-specific oligonucleotide sets that act as unique "genomic fingerprints".
• Achieved 92% accuracy in placing 100 kb DNA sequences to their correct species using the scoring algorithm.
• Demonstrated improved ability to relate phage genomes to their bacterial hosts compared to previous methods.

Conclusions:

• The discovered oligonucleotides provide unique genomic fingerprints for bacterial identification and classification.
• The approach offers a novel basis for phylogenetic studies.
• The methods are well-suited for classifying short DNA fragments from environmental shotgun sequencing.
• The developed algorithms have potential applications in broader bioinformatics challenges.