SeqNLS: nuclear localization signal prediction based on frequent pattern mining and linear motif scoring

Area of Science:

• Molecular Biology
• Bioinformatics
• Computational Biology

Background:

• Nuclear localization signals (NLSs) are crucial for protein import into the nucleus.
• Existing NLS motif knowledge is limited, hindering comprehensive identification.
• Diverse NLS patterns necessitate advanced computational approaches.

Purpose of the Study:

• To develop a novel algorithm, SeqNLS, for identifying potential NLS patterns.
• To overcome limitations of current NLS motif databases.
• To enhance the prediction accuracy of nuclear import signals.

Main Methods:

• Sequential pattern mining to extract frequent NLS patterns.
• Prediction of NLS candidates using identified patterns.
• Filtering candidates with a linear motif-scoring scheme incorporating sequence disorder and local conservation (IRLC).

Main Results:

• SeqNLS effectively detects potential NLSs on Yeast and Hybrid datasets.
• Linear motif features significantly complement sequence features for NLS discernment.
• SeqNLS achieves over 50% NLS identification with >=0.7 precision, outperforming state-of-the-art methods.

Conclusions:

• SeqNLS is a robust and effective tool for NLS prediction.
• The integration of sequence and linear motif features enhances NLS identification accuracy.
• SeqNLS offers improved performance over existing methods for predicting nuclear import signals.