QT Clustering using Euclidean Distance, Jackknife Correlation

QT Clustering using Euclidean Distance, Jackknife Correlation performs clustering of genome-wide gene expression data to identify stable co-expression groups using Euclidean distance and jackknife correlation for validation, implemented as a MATLAB-based analysis suite and demonstrated on yeast cell cycle studies.

Key Features:

• Similarity Measure: A specialized similarity measure aimed at minimizing false positives when grouping gene expression profiles.
• Clustering Algorithm: A custom clustering algorithm tailored to capture temporal and spatial dynamics in gene expression patterns for precise grouping.
• Euclidean Distance: Uses Euclidean distance metrics to quantify similarity between expression profiles.
• Jackknife Correlation Validation: Employs jackknife correlation to assess cluster stability and reliability by re-evaluating assignments with subsets excluded.
• Implementation: Implemented as a MATLAB-based analytical suite.

Scientific Applications:

• Genome-wide expression analysis: Summarizes and clusters large-scale gene expression datasets to reveal co-expression groups.
• Yeast cell cycle studies: Demonstrated utility in identifying biologically meaningful gene clusters in yeast cell cycle datasets.
• Downstream supervised analyses and inference: Clustered gene sets facilitate supervised analyses to investigate cellular processes, regulatory mechanisms, and potential genetic or pharmacological targets.

Methodology:

The method applies a similarity measure to reduce noise and false positives, groups genes using a custom clustering algorithm that leverages Euclidean distance, and assesses cluster stability using jackknife correlation by excluding subsets of data points; the suite is implemented in MATLAB.