Feedback GAN (FBGAN) for DNA: a Novel Feedback-Loop Architecture for Optimizing Protein Functions

By Anvita Gupta, James Y. Zou. 2018

62 citations recorded by the source corpus.

Generative Adversarial Networks (GANs) represent an attractive and novel approach to generate realistic data, such as genes, proteins, or drugs, in synthetic biology. Here, we apply GANs to generate synthetic DNA sequences encoding for proteins of variable length. We propose a novel feedback-loop architecture, called Feedback GAN (FBGAN), to optimize the synthetic gene sequences for desired properties using an external function analyzer. The proposed architecture also has the advantage that the analyzer need not be differentiable. We apply the feedback-loop mechanism to two examples: 1) generating synthetic genes coding for antimicrobial peptides, and 2) optimizing synthetic genes for the secondary structure of their resulting peptides. A suite of metrics demonstrate that the GAN generated proteins have desirable biophysical properties. The FBGAN architecture can also be used to optimize GAN-generated datapoints for useful properties in domains beyond genomics.

Subfields: Molecular Biology

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Molecular Biology

Citations in the corpus

Conditioning by adaptive sampling for robust design (2019, methodology): “(2) FBVAE— the method of Gupta and Zou (2018). This method does not require a differentiable oracle. (3) GB-NO— the approach described by Gómez-Bombarelli et al. (2018). This method requires a differentiable oracle. (4) GB —the approach implemented by Gómez-Bombarelli et al. (2018) which has some ad”
Machine learning in protein engineering (2018, methodology): “…and generative adverserial networks (Goodfellow et al., 2014) have been used to generate novel antimicrobial peptides (M¨uller et al., 2018; Gupta & Zou, 2018) and protein structures (Anand & Huang, 2018), and there has been an effort to develop a mathematical framework for using a…”
AdaLead: A simple and robust adaptive greedy search algorithm for sequence design (2020, background): “In Feedback-GAN (FBGAN), Gupta & Zou (2018) pair a generative adversarial network (Goodfellow et al., 2014) that is trained to predict whether sequences belong to the functional set using a frozen oracle φ (cid:48) which ﬁlters a subset of sequences at each training step.”
Machine learning applications in drug development (2019, unknown): “Another paper [52] uses, as ‘‘predictor‘‘ network, a NN which predicts the probability of the (generated) DNA sequence of coding for an antimicrobial peptide (AMP), and have succeeded in training a GAN which returns 77.”
A Review on Generative Adversarial Networks: Algorithms, Theory, and Applications (2020, methodology): “Medical field: GANs have been widely utilized in medical field such as generating and designing DNA [404], [405], drug discovery [406], generating multi-label discrete patient records [407], medical image processing [408]–[415], dental restorations [416], and doctor recommendation [417].”
De Novo Protein Design for Novel Folds using Guided Conditional Wasserstein Generative Adversarial Networks (gcWGAN) (2019, background): “, 2014) and Variational Auto-Encoder (VAE) (Kingma and Welling, 2013), have reached the sibling fields of inverse design for DNA (Gupta and Zou, 2018; Killoran et al., 2017), RNA (Eastman et al.”