N-gram analysis of 970 microbial organisms reveals presence of biological language models
2011

N-gram Analysis of Microbial Proteomes

Sample size: 970 publication Evidence: moderate

Author Information

Author(s): Osmanbeyoglu Hatice Ulku, Ganapathiraju Madhavi K

Primary Institution: University of Pittsburgh

Hypothesis

Do microbial proteome sequences exhibit characteristics similar to natural language texts, allowing for the identification of biological language models?

Conclusion

Microbial proteome sequences contain specific n-gram sequences that serve as signatures, and perplexity can predict evolutionary distances within genera.

Supporting Evidence

  • N-gram frequencies can reveal unique signatures for different microbial organisms.
  • Perplexity values correlate with evolutionary distances among related microbes.
  • The study analyzed 970 microbial proteomes, significantly expanding previous research.

Takeaway

Scientists studied the proteins of 970 different microbes to see if they have unique patterns, like how words are used in a language, and found that these patterns can help us understand how closely related different microbes are.

Methodology

Statistical n-gram analysis was applied to whole proteome sequences using the Biological Language Modeling Toolkit and Patternix Revelio toolkit.

Limitations

The study is limited to the available genome sequences and may not generalize to all microbial organisms.

Digital Object Identifier (DOI)

10.1186/1471-2105-12-12

Want to read the original?

Access the complete publication on the publisher's website

View Original Publication