Identify and analysis crotonylation sites in histone by using support vector machines

Area of Science:

• Biochemistry and Molecular Biology
• Genomics and Epigenetics
• Computational Biology

Background:

• Lysine crotonylation (Kcr) is a recently identified histone modification enriched at active gene regulatory elements in mammalian genomes.
• Current experimental methods for identifying Kcr sites, such as high-resolution mass spectrometry, are costly and time-consuming.
• There is a need for efficient computational tools to facilitate Kcr site identification.

Purpose of the Study:

• To develop and validate a novel computational method for predicting lysine crotonylation sites in histones.
• To establish an effective sequence information extraction scheme for histone modifications.
• To provide a cost-effective and rapid alternative to experimental Kcr site identification.

Main Methods:

• Proposed a new encoding scheme: position weight amino acid composition to extract histone sequence information around crotonylation sites.
• Constructed a rigorous benchmark dataset from Uniprot protein data for training and testing.
• Employed a support vector machine (SVM) classifier for predicting Kcr sites.

Main Results:

• The developed computational model achieved high performance metrics: 71.69% sensitivity, 98.7% specificity, 94.43% accuracy, and 0.778 MCC via jackknife cross-validation.
• The proposed model outperformed the random forest algorithm in Kcr site prediction.
• Feature analysis was conducted on the identified samples.

Conclusions:

• Accurate identification of Kcr sites is crucial for understanding histone function and regulatory roles.
• This computational method can significantly advance the understanding of crotonylation's physiological significance.
• The findings offer valuable insights for developing therapeutic strategies targeting crotonylation-related diseases.