APT: An Automated Probe Tracker From Gene Expression Data

Area of Science:

• Bioinformatics
• Computational Biology
• Genomics

Background:

• Semi-automatic tools like ArrayMining and GEO2R are popular for detecting diagnostic probes but have limitations.
• Existing tools rank probes by statistical significance, requiring user input or offering limited gene outputs.
• This necessitates a more automated and accurate method for diagnostic probe selection.

Purpose of the Study:

• To develop a fully automatic probe-set selection method.
• To improve the accuracy and efficiency of identifying diagnostic probes for pathophysiological conditions.
• To compare the performance of the new method against existing popular tools.

Main Methods:

• Developed an automatic probe-set selection tool named APT.
• Utilized a voting system combining outputs from t-Test, Mann-Whitney Test, and Empirical Bayes Moderated t-test.
• Demonstrated that statistical method parameters can be mathematically linked to group fisher's discriminant ratio.

Main Results:

• APT achieved an 88.97% success rate in identifying reported probes, surpassing ArrayMining (80.40%) and GEO2R (87.60%).
• The new method showed superior performance in sensitivity and specificity across 10-fold cross-validation and 5 new test cases.
• APT offers a fully automatic approach, eliminating the need for manual probe number input.

Conclusions:

• APT represents a significant advancement in automatic diagnostic probe selection.
• The method provides higher accuracy, sensitivity, and specificity compared to current popular tools.
• APT offers a more efficient and reliable approach for identifying disease-related diagnostic probes.