RNA-seq assistant: machine learning based methods to identify more transcriptional regulated genes.

Area of Science:

• Genomics
• Bioinformatics
• Computational Biology

Background:

• RNA-sequencing (RNA-seq) is crucial for gene expression analysis but faces limitations in detecting all differentially expressed genes (DEGs) due to experimental and analytical biases.
• Traditional RNA-seq workflows, including alignment, quantification, normalization, and statistical analysis, can lead to false positives and negatives.
• Machine learning (ML) offers a promising avenue to overcome these limitations by learning patterns from data for improved gene expression prediction.

Purpose of the Study:

• To investigate the potential of ML-based methods to enhance the identification of DEGs missed by conventional RNA-seq.
• To evaluate the performance of different ML algorithms and feature selection techniques for DEG prediction.

Main Methods:

• Assessed the performance of three feature selection algorithms combined with five classification methods on training and testing datasets.
• Identified the top 23 most informative features for DEG prediction.
• Utilized InfoGain feature selection and Logistic Regression for the optimal ML model.

Main Results:

• The study identified a powerful ML model combining InfoGain feature selection and Logistic Regression for predicting DEGs.
• The model's effectiveness was validated through predictions on ethylene-regulated gene expression.
• Quantitative reverse transcription PCR (qRT-PCR) confirmed the performance of the ML-based prediction.

Conclusions:

• Integrating ML methods with RNA-seq significantly boosts the sensitivity of DEG identification.
• This combined approach offers a more comprehensive and accurate analysis of gene expression dynamics.