Incorporating time-delays in S-System model for reverse engineering genetic networks

Area of Science:

• Systems Biology
• Computational Biology
• Bioinformatics

Background:

• Gene regulatory networks (GRNs) involve complex, instantaneous or time-delayed interactions between transcription factors and target genes.
• Existing GRN inference models often fail to simultaneously represent both interaction types, limiting their accuracy.
• The S-System model, while useful, is restricted to modeling only instantaneous interactions.

Purpose of the Study:

• To develop a novel modeling approach capable of simultaneously inferring both instantaneous and time-delayed interactions in GRNs.
• To enhance the accuracy and computational efficiency of GRN inference.
• To address the limitations of existing S-System models in capturing time-delayed regulatory events.

Main Methods:

• Introduction of a Time-Delayed S-System (TDSS) model utilizing delay differential equations.
• Incorporation of delay parameters that can be fractional, not just integer values.
• Development of a new model evaluation criterion leveraging the sparse and scale-free properties of GRNs to optimize search space and reduce computation time.

Main Results:

• The TDSS model successfully captures both instantaneous and time-delayed interactions with high precision in synthetic networks.
• A novel evaluation criterion significantly reduces computation time while improving model accuracy by adapting network properties.
• Fractional delay parameters offer greater flexibility in modeling system dynamics.

Conclusions:

• The proposed TDSS model accurately infers both instantaneous and time-delayed interactions in GRNs.
• Experimental validation on synthetic and real-world networks (IRMA, SOS DNA repair) demonstrates significant performance improvements over state-of-the-art methods.
• The novel evaluation criterion enhances the efficiency and accuracy of GRN modeling.