Deep Hidden Physics Modeling of Cell Signaling Networks.

Area of Science:

• Oncology
• Bioinformatics
• Computational Biology

Background:

• Cancer is a leading global cause of death, incurring substantial economic costs.
• Current cancer therapeutics face challenges including low market approval rates and drug resistance.
• Kinase signaling pathways are crucial in cancer, but existing inhibitors cause toxicity and resistance.

Purpose of the Study:

• To address the urgent need for novel cancer therapeutics.
• To improve the understanding and modeling of cancer-related cell signaling networks.
• To develop advanced computational models for predicting molecular signaling alterations in cancer.

Main Methods:

• Utilizing data-driven deep-learning approaches.
• Developing mechanistic computational models.
• Generating in silico probabilistic predictions of molecular signaling network rearrangements.

Main Results:

• The study proposes a framework for sophisticated computational modeling.
• The models aim to predict causal molecular signaling network alterations in cancer.
• This approach facilitates global and mechanistic modeling of cancer signaling.

Conclusions:

• Advanced computational models are essential for understanding cancer signaling.
• Data-driven and mechanistic models can improve the prediction of cancer-related molecular changes.
• This work supports the development of more effective and sustainable cancer therapeutics.