MHC-Binding Peptide Generation Using Dual-Projection GANs
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing computational systems are incapable of generating new binding peptides for Major Histocompatibility Complex (MHC) proteins given existing positive binding peptide examples, limiting the efficiency of clinical peptide vaccine search and design.
Innovation Solution
A Generative Adversarial Network (GAN) is trained on a dataset of binding and non-binding peptide sequences, using dual projection vectors and cross-entropy losses to generate new binding peptides, with a discriminator distinguishing between real and generated sequences and a generator fooling the discriminator.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If traditional computational systems are used to predict peptide-MHC binding interactions, then binding prediction can be performed, but new binding peptides cannot be generated from existing positive examples
Solution Approach 1:
The GAN generator creates copies of binding peptides by learning the underlying distribution of positive binding examples and generating new sequences that follow the same patterns. The generator iteratively produces peptide sequences that mimic the characteristics of known binders, enabling expansion of the binding peptide repertoire from limited positive examples.
Solution Approach 2:
The discriminator provides feedback to the generator by distinguishing real binding peptides from generated ones. This adversarial feedback loop allows the generator to progressively improve its peptide generation capability, learning from the discriminator's ability to identify authentic binders and adjusting its output to better match the binding peptide distribution.
2Productivity
If GAN training is performed only on binding peptide sequences, then generation capability is improved, but training complexity increases with dual projection vectors and cross-entropy losses
Solution Approach 1:
The training objective is segmented into two distinct cross-entropy losses: one for the generator to distinguish binding from non-binding peptides, and another for the discriminator to differentiate real binding peptides from generated ones. This segmentation allows each component to be optimized independently for its specific function while working toward the common goal of improved peptide generation.
Solution Approach 2:
The model extends the standard GAN framework by adding dual projection vectors that map peptides into a latent space optimized for binding prediction. This dimensional extension allows the model to capture complex binding patterns while maintaining a structured training approach through the additional projection layers.
Data Source
AI summary
A method generates new binding peptides to Major Histocompatibility Complex (MHC) proteins. The method includes training, a Generative Adversarial Network (GAN) having a generator and a discriminator only on a set of binding peptide sequences given training data comprising the set of binding peptide sequences and a set of non-binding peptide sequences. A GAN training objective includes the discriminator being iteratively updated to distinguish generated peptide sequences from sampled binding peptide sequences as fake or real and the generator being updated to fool the discriminator. The GAN training objective is optimized while learning two projection vectors for a binding class with two cross-entropy losses. A first loss discriminates binding peptide sequences in the training data from non-binding peptide sequences in the training data. A second loss discriminates generated binding peptide sequences from non-binding peptide sequences in the training data. The method enables supporting decision-making related to new binding peptide generation.