DL-CRISPR: A Deep Learning Method for Off-Target Activity Prediction in CRISPR/Cas9 With Data Augmentation

Abstract

Clustered Regularly Interspaced Short Palindromic Repeats (CRISPR)/CRISPR-associated (Cas) system is a popular and easy to use gene-editing technique, but it has off-target risk. Cutting the off-target sites will harm the cells severely, hence in silico methods are needed to help to avoid this. Most existing in silico approaches mainly relied on a relatively small positive dataset and the data imbalance issue still exists. Besides, some samples used to be considered as negative are later proved to be positive. Hence, it is essential to refresh the dataset and develop more accurate off-target activity prediction programs. In this work, firstly, we extended the current positive dataset and explored the potential differences between positive and negative data based on the new dataset. Then we adopted a new data augmentation method to solve the data imbalance issue, and used the ensemble idea to take more negative data into consideration to make the model close to the real scenario, but at the same time keeping the model balance. Finally, we developed DL-CRISPR, a deep learning framework to predict off-target activity in CRISPR/Cas9. DL-CRISPR is evaluated and compared with other state-of-the-art methods on three kinds of datasets: 5-fold cross validation test datasets, putative off-targets datasets related to specific single guide RNAs (sgRNAs), and putative off-targets datasets related to unseen sgRNAs. DL-CRISPR realizes the best average accuracy, i.e. 98.57%, on 5-fold cross validation datasets and correctly detects more off-targets on datasets related to both seen and unseen sgRNAs. INDEX TERMS CRISPR/Cas9, data augmentation, deep learning, off-target.

R-CRISPR: A Deep Learning Network to Predict Off-Target Activities with Mismatch, Insertion and Deletion in CRISPR-Cas9 System

“…Seven off-target datasets that were validated by mainstream experimental methods were selected for model training and validation [ 31 ]. Those datasets were shifted into two categories: one category contains mismatch, insertion, and deletion off-target sites, while another just includes datasets with mismatch off-target sites.…”

“…Off-target prediction model R-CRISPR was inspired by the LRCN framework and includes an encoding scheme [ 31 ] to convert the on- and off-target pair into suitable input for neural network, a convolutional layer, and a recurrent layer. The convolutional layer built on the architecture of CNN and RepVGG [ 39 ] module is used as a feature extractor, while the recurrent layer is composed of bi-directional LSTM RNN, and the output of the recurrent layer is passed to the subsequent dense layers.…”

“…This method also included a replacement-based sensitivity analysis to illustrate the relative importance of each site. Different from those methods that improved on model architecture, DL-CRISPR [ 31 ] focused on dataset optimization. They extended the current positive dataset to improve the competitiveness of the model and investigated dataset design to address data imbalance issue, after that, four layers of CNN were used to learn data features and the final score is got as the average score of 10 models.…”

Examining Machine Learning Models That Predict sgRNA Cleavage Efficiencies

“…The source code for this learning model is located on github: https://github.com/MoonLBH/CNN-XG DL-CRISPR first had to obtain a lot of data. Scientists "collected off-target sequences as well as their sgRNAs from in vitro -and cell-based genome-wide assays" (Zhang et. al., 2020).…”

A systematic evaluation of data processing and problem formulation of CRISPR off-target site prediction

“…These methods can be split into three categories: alignment-based, ruled-based, and data-driven-based. Many data-driven methods are based on machine learning, such as CRISTA, DeepCrispr, Elevation, among many others [25, 26, 27, 28, 29, 30, 31, 32]. Still, all these methods were trained using relatively small experimental data, i.e.…”