Search

The vast expansion of protein sequence databases provides an opportunity for new protein design approaches which seek to learn the sequence-function relationship directly from natural sequence variation. Deep generative models trained on protein sequence data have been shown to learn biologically meaningful representations helpful for a variety of downstream tasks, but their potential for direct use in the design of novel proteins remains largely unexplored. Here we show that variational autoencoders trained on a dataset of almost 70000 luciferase-like oxidoreductases can be used to generate novel, functional variants of the luxA bacterial luciferase. We propose separate VAE models to work with aligned sequence input (MSA VAE) and raw sequence input (AR-VAE), and offer evidence that while both are able to reproduce patterns of amino acid usage characteristic of the family, the MSA VAE is better able to capture long-distance dependencies reflecting the influence of 3D structure. To confirm the practical utility of the models, we used them to generate variants of luxA whose luminescence activity was validated experimentally. We further showed that conditional variants of both models could be used to increase the solubility of luxA without disrupting function. Altogether 6/12 of the variants generated using the unconditional AR-VAE and 9/11 generated using the unconditional MSA VAE retained measurable luminescence, together with all 23 of the less distant variants generated by conditional versions of the models; the most distant functional variant contained 35 differences relative to the nearest training set sequence. These results demonstrate the feasibility of using deep generative models to explore the space of possible protein sequences and generate useful variants, providing a method complementary to rational design and directed evolution approaches., Author summary The design of novel proteins with specified function and biochemical properties is a longstanding goal in bio-engineering with applications across medicine and nanotechnology. Despite the impressive achievements of traditional approaches, a great deal of scope remains for the development of data-driven methods capable of exploiting the record of natural sequence variation available in protein databases. Deep generative models such as variational autoencoders (VAEs) have shown remarkable success in synthesising realistic data samples across a range of modalities, driving recent interest in developing such models for proteins. However, experimental evidence for the viability of such techniques in practical protein design settings remains scarce. Here we show that VAEs trained on the family of luciferase-like oxidoreductases can be used to generate functional variants of the luxA bacterial luciferase. We compare the use of raw and aligned sequences as input to the model, providing evidence that models trained on aligned data are better able to learn functional constraints. Finally, we demonstrate the possibility of controlling desired properties of the designed sequences, by using conditional versions of the VAE models to increase the solubility of the wild-type luxA sequence from P. luminescens.

With the increasing application of deep learning methods to the modelling of regulatory DNA sequences has come an interest in exploring what types of architecture are best suited to the domain. Networks designed to predict many functional characteristics of noncoding DNA in a multitask framework have to recognise a large number of motifs and as a result benefit from large numbers of convolutional filters in the first layer. The use of large first layers in turn motivates an exploration of strategies for addressing the sparsity of output and possibility for overfitting that result. To this end we propose the use of a dimensionality-reducing linear projection layer after the initial motif-recognising convolutions. In experiments with a reduced version of the DeepSEA dataset we find that inserting this layer in combination with dropout into convolutional and convolutional-recurrent architectures can improve predictive performance across a range of first layer sizes. We further validate our approach by incorporating the projection layer into a new convolutional-recurrent architecture which achieves state of the art performance on the full DeepSEA dataset. Analysis of the learned projection weights shows that the inclusion of this layer simplifies the network’s internal representation of the occurrence of motifs, notably by projecting features representing forward and reverse-complement motifs to similar positions in the lower dimensional feature space output by the layer.

With the increasing application of deep learning methods to the modelling of regulatory DNA sequences has come an interest in exploring what types of architecture are best suited to the domain. Networks designed to predict many functional characteristics of noncoding DNA in a multitask framework have to recognise a large number of motifs and as a result benefit from large numbers of convolutional filters in the first layer. The use of large first layers in turn motivates an exploration of strategies for addressing the sparsity of output and possibility for overfitting that result. To this end we propose the use of a dimensionality-reducing linear projection layer after the initial motif-recognising convolutions. In experiments with a reduced version of the DeepSEA dataset we find that inserting this layer in combination with dropout into convolutional and convolutional-recurrent architectures can improve predictive performance across a range of first layer sizes. We further validate our approach by incorporating the projection layer into a new convolutional-recurrent architecture which achieves state of the art performance on the full DeepSEA dataset. Analysis of the learned projection weights shows that the inclusion of this layer simplifies the network’s internal representation of the occurrence of motifs, notably by projecting features representing forward and reverse-complement motifs to similar positions in the lower dimensional feature space output by the layer.