Identifying microRNAs involved in cancer pathway using support vector machines

Area of Science:

• Biochemistry
• Bioinformatics
• Genomics

Background:

• Small non-protein coding RNAs, including microRNAs (miRs), play diverse biological roles.
• miRs are increasingly recognized for their association with cancer, driving interest in their therapeutic applications.
• Experimental identification of cancer-associated miRs is a complex and lengthy process.

Purpose of the Study:

• To develop an in-silico method for identifying cancer-associated microRNAs (miRs).
• To accelerate the process of discovering miRs linked to cancer through computational analysis.

Main Methods:

• An in-silico study identified global signatures in experimentally validated cancer-associated miRs.
• A two-step binary classifier system (miRSEQ and miRINT) was developed using support vector machines.
• Sixty distinguishing features were extracted from miR sequences, folding thermodynamics, and miRNA-mRNA structures.

Main Results:

• The classifier system achieved good performance, with Matthew's correlation coefficient (MCC) values ranging from 0.72 to 0.82.
• The developed computational tool aids in the identification of cancer-associated miRs.

Conclusions:

• The developed in-silico approach offers an efficient method for identifying cancer-associated miRs.
• This tool can expedite research into the role of miRs in cancer and their therapeutic potential.