DeepBIO: an automated and interpretable deep-learning platform for high-throughput biological sequence prediction, functional annotation and visualization analysis

Abstract

Here, we present DeepBIO, the first-of-its-kind automated and interpretable deep-learning platform for high-throughput biological sequence functional analysis. DeepBIO is a one-stop-shop web service that enables researchers to develop new deep-learning architectures to answer any biological question. Specifically, given any biological sequence data, DeepBIO supports a total of 42 state-of-the-art deep-learning algorithms for model training, comparison, optimization and evaluation in a fully automated pipeline. DeepBIO provides a comprehensive result visualization analysis for predictive models covering several aspects, such as model interpretability, feature analysis and functional sequential region discovery. Additionally, DeepBIO supports nine base-level functional annotation tasks using deep-learning architectures, with comprehensive interpretations and graphical visualizations to validate the reliability of annotated sites. Empowered by high-performance computers, DeepBIO allows ultra-fast prediction with up to million-scale sequence data in a few hours, demonstrating its usability in real application scenarios. Case study results show that DeepBIO provides an accurate, robust and interpretable prediction, demonstrating the power of deep learning in biological sequence functional analysis. Overall, we expect DeepBIO to ensure the reproducibility of deep-learning biological sequence analysis, lessen the programming and hardware burden for biologists and provide meaningful functional insights at both the sequence level and base level from biological sequences alone. DeepBIO is publicly available at https://inner.wei-group.net/DeepBIO.

Multi-CGAN: Deep Generative Model-Based Multiproperty Antimicrobial Peptide Design

“…We further explored the applicability of Multi-CGAN for data augmentation, which may aid in alleviating the few sample learning problems of some downstream tasks. In particular, we employed a deep learning platform named DeepBIO to conduct the experiments . DeepBIO can provide users with a total of 42 state-of-the-art deep learning methods for model training, comparison, optimization, and evaluation in a fully automated pipeline.…”

“…In particular, we employed a deep learning platform named DeepBIO to conduct the experiments. 38 DeepBIO can provide users with a total of 42 state-of-the-art deep learning methods for model training, comparison, optimization, and evaluation in a fully automated pipeline. Here, we chose the AMPs data set constructed by Zhang et al as our downstream data set for the binary classification task and selected several classical models (DNN, LSTM, VDCNN, Transformer) on DeepBIO as the baseline methods.…”

Engineering extracellular electron transfer pathways of electroactive microorganisms by synthetic biology for energy and chemicals production

“…46 The future studies of MESs may focus on the following aspects: (i) combining genomic analysis, microbial physiological, biochemical analysis, and spectroscopy analysis to investigate the mechanisms of cathode-cells interfacial electron uptake and energy conversion, thus facilitating rational engineering electrotrophs to increase chemicals synthesis efficiency in MESs. 583 (ii) Constructing artificial electrophobic channels assembled from conductive polymers, artificial nanowires, and three-dimensional conductive hybrid biofilms to facilitate efficient electron uptake across cell membranes. (iii) Combining genomics, deep learning, and laboratory evolutionary strategies to explore the design of new carbon/nitrogen fixation pathways or rationally optimize carbon flows to reduce the energy consumption of intracellular enzymatic reactions.…”

MIBPred: Ensemble Learning-Based Metal Ion-Binding Protein Classifier

“…In the past, many machine learning methods used for predicting metal ion-binding proteins have shown outstanding performance. For example, experts have developed MetalPredator and RPCIBP classifiers for predicting proteins related to iron and sulfur clusters and copper ion-binding proteins, respectively. − In addition, some studies have also achieved favorable results in predicting metal ion-binding sites, including iron ion-binding sites, zinc ion-binding sites, and copper ion-binding sites. − However, there is currently no efficient and stable multimetal ion-binding protein classifier.…”