Multi-target prediction for dummies using two-branch neural networks

Abstract Analyzing the relation between intelligence and neural activity is of the utmost importance in understanding the working principles of the human brain in health and disease. In existing literature, functional brain connectomes have been used successfully to predict cognitive measures such as intelligence quotient (IQ) scores in both healthy and disordered cohorts using machine learning models. However, existing methods resort to flattening the brain connectome (i.e., graph) through vectorization which overlooks its topological properties. To address this limitation and inspired from the emerging graph neural networks (GNNs), we design a novel regression GNN model (namely RegGNN) for predicting IQ scores from brain connectivity. On top of that, we introduce a novel, fully modular sample selection method to select the best samples to learn from for our target prediction task. However, since such deep learning architectures are computationally expensive to train, we further propose a \emph{learning-based sample selection} method that learns how to choose the training samples with the highest expected predictive power on unseen samples. For this, we capitalize on the fact that connectomes (i.e., their adjacency matrices) lie in the symmetric positive definite (SPD) matrix cone. Our results on full-scale and verbal IQ prediction outperforms comparison methods in autism spectrum disorder cohorts and achieves a competitive performance for neurotypical subjects using 3-fold cross-validation. Furthermore, we show that our sample selection approach generalizes to other learning-based methods, which shows its usefulness beyond our GNN architecture.

Download Full-text

miTAR: a hybrid deep learning-based approach for predicting miRNA targets

BMC Bioinformatics ◽

10.1186/s12859-021-04026-6 ◽

2021 ◽

Vol 22 (1) ◽

Author(s):

Tongjun Gu ◽

Xiwu Zhao ◽

William Bradley Barbazuk ◽

Ji-Hyun Lee

Keyword(s):

Neural Networks ◽

Deep Learning ◽

Mirna Target ◽

Target Prediction ◽

Hybrid Models ◽

Mirna Target Prediction ◽

Max Pooling ◽

Mirna Targets ◽

Spatial Features ◽

Wide Range

Abstract Background microRNAs (miRNAs) have been shown to play essential roles in a wide range of biological processes. Many computational methods have been developed to identify targets of miRNAs. However, the majority of these methods depend on pre-defined features that require considerable efforts and resources to compute and often prove suboptimal at predicting miRNA targets. Results We developed a novel hybrid deep learning-based (DL-based) approach that is capable of predicting miRNA targets at a higher accuracy. This approach integrates convolutional neural networks (CNNs) that excel in learning spatial features and recurrent neural networks (RNNs) that discern sequential features. Therefore, our approach has the advantages of learning both the intrinsic spatial and sequential features of miRNA:target. The inputs for our approach are raw sequences of miRNAs and genes that can be obtained effortlessly. We applied our approach on two human datasets from recently miRNA target prediction studies and trained two models. We demonstrated that the two models consistently outperform the previous methods according to evaluation metrics on test datasets. Comparing our approach with currently available alternatives on independent datasets shows that our approach delivers substantial improvements in performance. We also show with multiple evidences that our approach is more robust than other methods on small datasets. Our study is the first study to perform comparisons across multiple existing DL-based approaches on miRNA target prediction. Furthermore, we examined the contribution of a Max pooling layer in between the CNN and RNN and demonstrated that it improves the performance of all our models. Finally, a unified model was developed that is robust on fitting different input datasets. Conclusions We present a new DL-based approach for predicting miRNA targets and demonstrate that our approach outperforms the current alternatives. We supplied an easy-to-use tool, miTAR, at https://github.com/tjgu/miTAR. Furthermore, our analysis results support that Max Pooling generally benefits the hybrid models and potentially prevents overfitting for hybrid models.

Download Full-text

Accurate deep learning off-target prediction with novel sgRNA-DNA sequence encoding in CRISPR-Cas9 gene editing

Bioinformatics ◽

10.1093/bioinformatics/btab112 ◽

2021 ◽

Author(s):

Jeremy Charlier ◽

Robert Nadon ◽

Vladimir Makarenkov

Keyword(s):

Neural Networks ◽

Deep Learning ◽

Dna Sequence ◽

Dna Sequences ◽

Gene Editing ◽

Sequence Data ◽

Target Prediction ◽

Feedforward Neural Networks ◽

Strong Impact ◽

Sequence Encoding

Abstract Motivation Off-target predictions are crucial in gene editing research. Recently, significant progress has been made in the field of prediction of off-target mutations, particularly with CRISPR-Cas9 data, thanks to the use of deep learning. CRISPR-Cas9 is a gene editing technique which allows manipulation of DNA fragments. The sgRNA-DNA (single guide RNA-DNA) sequence encoding for deep neural networks, however, has a strong impact on the prediction accuracy. We propose a novel encoding of sgRNA-DNA sequences that aggregates sequence data with no loss of information. Results In our experiments, we compare the proposed sgRNA-DNA sequence encoding applied in a deep learning prediction framework with state-of-the-art encoding and prediction methods. We demonstrate the superior accuracy of our approach in a simulation study involving Feedforward Neural Networks (FNNs), Convolutional Neural Networks (CNNs) and Recurrent Neural Networks (RNNs) as well as the traditional Random Forest (RF), Naive Bayes (NB) and Logistic Regression (LR) classifiers.We highlight the quality of our results by building several FNNs, CNNs and RNNs with various layer depths and performing predictions on two popular CRISPOR and GUIDE-seq gene editing data sets. In all our experiments, the new encoding led to more accurate off-target prediction results, providing an improvement of the area under the Receiver Operating Characteristic (ROC) curve up to 35%. Availability The code and data used in this study are available at: https://github.com/dagrate/dl-offtarget

Download Full-text

miTAR: a hybrid deep learning-based approach for predicting miRNA targets

10.1101/2020.04.02.022608 ◽

2020 ◽

Author(s):

Tongjun Gu ◽

Xiwu Zhao ◽

William Bradley Barbazuk ◽

Ji-Hyun Lee

Keyword(s):

Neural Networks ◽

Deep Learning ◽

Mirna Target ◽

Target Prediction ◽

Unified Model ◽

Brier Score ◽

Mirna Target Prediction ◽

Learning Methods ◽

Mirna Targets ◽

Wide Range

AbstractmicroRNAs (miRNAs) are a major type of small RNA that alter gene expression at the post-transcriptional or translational level. They have been shown to play important roles in a wide range of biological processes. Many computational methods have been developed to predict targets of miRNAs in order to understand miRNAs’ function. However, the majority of the methods depend on a set of pre-defined features that require considerable effort and resources to compute, and these methods often do not effectively on the prediction of miRNA targets. Therefore, we developed a novel hybrid deep learning-based approach that is capable to predict miRNA targets at a higher accuracy. Our approach integrates two deep learning methods: convolutional neural networks (CNNs) that excel in learning spatial features, and recurrent neural networks (RNNs) that discern sequential features. By combining CNNs and RNNs, our approach has the advantages of learning both the intrinsic spatial and sequential features of miRNA:target. The inputs for the approach are raw sequences of miRNA and gene sequences. Data from two latest miRNA target prediction studies were used in our study: the DeepMirTar dataset and the miRAW dataset. Two models were obtained by training on the two datasets separately. The models achieved a higher accuracy than the methods developed in the previous studies: 0.9787 vs. 0.9348 for the DeepMirTar dataset; 0.9649 vs. 0.935 for the miRAW dataset. We also calculated a series of model evaluation metrics including sensitivity, specificity, F-score and Brier Score. Our approach consistently outperformed the current methods. In addition, we compared our approach with earlier developed deep learning methods, resulting in an overall better performance. Lastly, a unified model for both datasets was developed with an accuracy higher than the current methods (0.9545). We named the unified model miTAR for miRNA target prediction. The source code and executable are available at https://github.com/tjgu/miTAR.

Download Full-text

Using deep bidirectional recurrent neural networks for prosodic-target prediction in a unit-selection text-to-speech system

10.21437/interspeech.2015-372 ◽

2015 ◽

Author(s):

Raul Fernandez ◽

Asaf Rendel ◽

Bhuvana Ramabhadran ◽

Ron Hoory

Keyword(s):

Neural Networks ◽

Recurrent Neural Networks ◽

Target Prediction ◽

Text To Speech ◽

Unit Selection

Download Full-text

Predicting cognitive scores with graph neural networks through sample selection learning

Brain Imaging and Behavior ◽

10.1007/s11682-021-00585-7 ◽

2021 ◽

Author(s):

Martin Hanik ◽

Mehmet Arif Demirtaş ◽

Mohammed Amine Gharsallaoui ◽

Islem Rekik

Keyword(s):

Neural Networks ◽

Brain Connectivity ◽

Sample Selection ◽

Target Prediction ◽

Autism Spectrum ◽

Brain Connectome ◽

Training Samples ◽

Health And Disease ◽

Iq Scores ◽

Graph Neural Networks

AbstractAnalyzing the relation between intelligence and neural activity is of the utmost importance in understanding the working principles of the human brain in health and disease. In existing literature, functional brain connectomes have been used successfully to predict cognitive measures such as intelligence quotient (IQ) scores in both healthy and disordered cohorts using machine learning models. However, existing methods resort to flattening the brain connectome (i.e., graph) through vectorization which overlooks its topological properties. To address this limitation and inspired from the emerging graph neural networks (GNNs), we design a novel regression GNN model (namely RegGNN) for predicting IQ scores from brain connectivity. On top of that, we introduce a novel, fully modular sample selection method to select the best samples to learn from for our target prediction task. However, since such deep learning architectures are computationally expensive to train, we further propose a learning-based sample selection method that learns how to choose the training samples with the highest expected predictive power on unseen samples. For this, we capitalize on the fact that connectomes (i.e., their adjacency matrices) lie in the symmetric positive definite (SPD) matrix cone. Our results on full-scale and verbal IQ prediction outperforms comparison methods in autism spectrum disorder cohorts and achieves a competitive performance for neurotypical subjects using 3-fold cross-validation. Furthermore, we show that our sample selection approach generalizes to other learning-based methods, which shows its usefulness beyond our GNN architecture.

Download Full-text