Predicting non-coding RNA genes in Escherichia coli with boosted genetic programming.

Area of Science:

• Genomics
• Molecular Biology
• Bioinformatics

Background:

• Escherichia coli (E.coli) possesses numerous non-coding RNA (ncRNA) genes, excluding transfer RNAs (tRNAs) and ribosomal RNAs (rRNAs).
• Existing prediction methods have identified over a thousand potential ncRNA genes, but only 95 have been validated by multiple studies.
• A significant gap exists in comprehensively identifying and validating ncRNA genes in E.coli.

Purpose of the Study:

• To introduce a novel computational method for predicting ncRNA genes in E.coli.
• To identify novel ncRNA candidates beyond currently known and previously predicted genes.
• To experimentally validate a subset of the predicted ncRNA candidates.

Main Methods:

• Development of a new method utilizing automatic discovery of sequence patterns for ncRNA gene prediction.
• Application of the method to the E.coli genome.
• Experimental validation of selected predicted ncRNA candidates through biological assays.

Main Results:

• The method predicted 135 novel ncRNA candidates.
• An additional 152 predicted genes overlapped with existing literature predictions.
• Experimental testing of 16 predictions confirmed 12 as actual ncRNA transcripts, with 6 of these being novel predictions.

Conclusions:

• The novel sequence pattern discovery method is effective for predicting ncRNA genes in E.coli.
• A high experimental confirmation rate suggests many untested predictions are also ncRNAs.
• E.coli is likely to harbor a greater number of ncRNA genes than currently estimated.