Rank-preserving biclustering algorithm: a case study on miRNA breast cancer

Area of Science:

• Bioinformatics
• Genomics
• Cancer Research

Background:

• MicroRNA (miRNA) expression alterations are implicated in cancer development.
• Effective biomarkers are crucial for early breast cancer diagnosis and monitoring.
• Identifying novel biomarkers can improve clinical treatment and diagnosis decisions.

Purpose of the Study:

• To develop and validate a novel biclustering algorithm for breast cancer miRNA analysis.
• To discover significant miRNA biomarkers associated with breast cancer.
• To enhance understanding of breast cancer through miRNA expression patterns.

Main Methods:

• Proposed a pattern-based parallel biclustering algorithm named Rank-Preserving Biclustering (RPBic).
• Utilized a modified all substrings common subsequence (ALCS) framework to identify rank-preserved rows.
• Applied RPBic to synthetic and real-world breast cancer miRNA datasets.

Main Results:

• RPBic demonstrated superior performance over existing algorithms on synthetic data.
• Identified 68 biclusters in breast cancer data with significant clinical characteristics.
• Discovered frequency-based (hsa-miR-410, hsa-miR-483-5p) and network-based (hsa-miR-454, hsa-miR-137) miRNA biomarkers strongly linked to breast cancer.

Conclusions:

• RPBic is an effective algorithm for identifying biologically relevant biclusters and miRNA biomarkers.
• The identified biomarkers have strong clinical relevance and connectivity with breast cancer.
• This research provides valuable insights for breast cancer diagnosis and treatment strategies.