PepCVAE: Semi-Supervised Targeted Design of Antimicrobial Peptide Sequences

Das, Payel; Wadhawan, Kahini; Chang, Oscar; Sercu, Tom; Santos, Cicero dos; Riemer, Matthew; Chenthamarakshan, Vijil; Padhi, Inkit; Mojsilovic, Aleksandra

doi:10.48550/arxiv.1810.07743

Cited by 22 publications

(36 citation statements)

References 25 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Third, leveraging a Gumbel-Softmax approximation [32] enabled a continuous approximation of sampling in the discrete space of amino acid sequences and thus a direct optimization of peptides generated by the model. Before, such optimizations required a complex multi-stage training [28]. Finally, HydrAMP is the only model controls in a parametrized way the model creativity understood as the number of modifications introduced to the query peptide.…”

Section: Discussionmentioning

confidence: 99%

“…2.2 HydrAMP outperforms other models in the ability to generate antimicrobial and highly active peptide candidates HydrAMP was compared to two alternative models: Basic and PepCVAE [28]. Basic is a standard cVAE, while PepC-VAE is a state-of-the-art approach to peptide generation using the conditional variational autoencoder framework.…”

Section: Hydramp -A Conditional Generative Model Of Peptide Sequencesmentioning

confidence: 99%

See 1 more Smart Citation

Discovering highly potent antimicrobial peptides with deep generative model HydrAMP

Szymczak

Możejko

Grzegorzek

et al. 2022

Preprint

View full text Add to dashboard Cite

Antimicrobial peptides (AMP) emerge as compounds that can alleviate the global health hazard of antimicrobial resistance. Since the repertoire of experimentally verified AMPs is limited, there is a need for novel computational approaches to peptide generation. For such approaches, exploring the amino-acid peptide representation space is infeasible due to its sparsity and combinatorial complexity. Thus, we propose HydrAMP, a conditional variational autoencoder that learns a lower-dimensional and continuous space of peptides' representations and captures their antimicrobial properties. HydrAMP outperforms other approaches in generating peptides, either de novo, or by analogue discovery, and leverages parameter-controlled creativity. The model disentangles the latent representation of a peptide from its antimicrobial conditions, allowing for targeted generation. Wet-lab experiments and molecular dynamics simulation confirm the increased activity of a Pexiganan-based analogue produced by HydrAMP. HydrAMP proposes new promising AMP candidates, enabling progress towards a new generation of antibiotics.

show abstract

Section: Discussionmentioning

confidence: 99%

Section: Hydramp -A Conditional Generative Model Of Peptide Sequencesmentioning

confidence: 99%

Discovering highly potent antimicrobial peptides with deep generative model HydrAMP

Szymczak

Możejko

Grzegorzek

et al. 2022

Preprint

View full text Add to dashboard Cite

show abstract

“…and literature [21], as well as scientific domains such as drug discovery [24], materials discovery [98], and software engineering [69]. Some of this work has examined the extent to which deep generative models provide humans with augmented capabilities, addressing the question of the extent to which the human-AI team produces outcomes better than those produced by either party alone (e.g.…”

Section: Co-creation With Generative Aimentioning

confidence: 99%

Perfection Not Required? Human-AI Partnerships in Code Translation

Weisz¹,

Müller²,

Houde³

et al. 2021

26th International Conference on Intelligent User Interfaces

View full text Add to dashboard Cite

Generative models have become adept at producing artifacts such as images, videos, and prose at human-like levels of proficiency. New generative techniques, such as unsupervised neural machine translation (NMT), have recently been applied to the task of generating source code, translating it from one programming language to another. The artifacts produced in this way may contain imperfections, such as compilation or logical errors. We examine the extent to which software engineers would tolerate such imperfections and explore ways to aid the detection and correction of those errors. Using a design scenario approach, we interviewed 11 software engineers to understand their reactions to the use of an NMT model in the context of application modernization, focusing on the task of translating source code from one language to another. Our threestage scenario sparked discussions about the utility and desirability of working with an imperfect AI system, how acceptance of that system's outputs would be established, and future opportunities for generative AI in application modernization. Our study highlights how UI features such as confidence highlighting and alternate translations help software engineers work with and better understand generative NMT models. CCS CONCEPTS• Human-centered computing → HCI theory, concepts and models; • Software and its engineering → Designing software; • Computing methodologies → Generative and developmental approaches.

show abstract

“…Moreover, Chen et al ( 2018) used joint embedding learning for text to shape generation. Joint sequencelabel embedding is also explored for or applied to molecular prediction/generation (Cao & Shen, 2021;Das et al, 2018).…”

Section: Related Workmentioning

confidence: 99%

Fold2Seq: A Joint Sequence(1D)-Fold(3D) Embedding-based Generative Model for Protein Design

Cao,

Das,

Chenthamarakshan

et al. 2021

Preprint

Self Cite

View full text Add to dashboard Cite

Designing novel protein sequences for a desired 3D topological fold is a fundamental yet nontrivial task in protein engineering. Challenges exist due to the complex sequence-fold relationship, as well as the difficulties to capture the diversity of the sequences (therefore structures and functions) within a fold. To overcome these challenges, we propose Fold2Seq, a novel transformer-based generative framework for designing protein sequences conditioned on a specific target fold. To model the complex sequence-structure relationship, Fold2Seq jointly learns a sequence embedding using a transformer and a fold embedding from the density of secondary structural elements in 3D voxels. On test sets with single, high-resolution and complete structure inputs for individual folds, our experiments demonstrate improved or comparable performance of Fold2Seq in terms of speed, coverage, and reliability for sequence design, when compared to existing state-of-the-art methods that include datadriven deep generative models and physics-based RosettaDesign. The unique advantages of foldbased Fold2Seq, in comparison to a structurebased deep model and RosettaDesign, become more evident on three additional real-world challenges originating from low-quality, incomplete, or ambiguous input structures. Source code and data are available at https://github.com/ IBM/fold2seq.

show abstract

PepCVAE: Semi-Supervised Targeted Design of Antimicrobial Peptide Sequences

Cited by 22 publications

References 25 publications

Discovering highly potent antimicrobial peptides with deep generative model HydrAMP

Discovering highly potent antimicrobial peptides with deep generative model HydrAMP

Perfection Not Required? Human-AI Partnerships in Code Translation

Fold2Seq: A Joint Sequence(1D)-Fold(3D) Embedding-based Generative Model for Protein Design

Contact Info

Product

Resources

About