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

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

A quantitative model for the dynamics of target recognition and off-target rejection by the CRISPR-Cas Cascade complex

“…The development of engineered effector variants (Amrani et al, 2018;Chen et al, 2017;Edraki et al, 2019;Gleditzsch et al, 2016;Kleinstiver et al, 2016;Lee et al, 2018;Luo et al, 2016;Ran et al, 2015;Slaymaker et al, 2016;Songailiene et al, 2019;Tuminauskaite et al, 2020;Wu et al, 2018) could recently reduce but not abolish off-targeting (Frock et al, 2015;Slaymaker et al, 2016). A frequently used complementary approach to prevent off-targets are in silico off-target predictors that promise to identify crRNAs with least promiscuity (Aprilyanto et al, 2021;Bae et al, 2014;Charlier et al, 2021;Haeussler et al, 2016;Lei et al, 2014;Lin and Wong, 2018;Minkenberg et al, 2019;Singh et al, 2015;Stemmer et al, 2015;Xu et al, 2017). Such prediction tools use heuristic scoring functions that try to reproduce sequence and mismatch position patterns from high throughput studies.…”

Effective use of sequence information to predict CRISPR-Cas9 off-target

“…To compare coding schemes more objectively, we use Dense Neural Networks (DNN), Convolutional Neural Networks (CNNs), and Recurrent Neural Networks (RNNs), three basic neural networks for evaluation. These three types of networks have been widely used in off-target prediction [20] . We built several different DNN, CNN, and RNN networks and evaluated them on two data sets to measure our coding scheme and Lin’s coding scheme.…”

“…Lin et al coded gRNA and DNA into a 4-dimensional one-hot vector and obtained the corresponding base pair code through the 'OR' operation [19] . Charlier et al pointed out that the 'OR' operation of the coding scheme would lead to information loss and proposed using the concatenating operation instead of the 'OR' operation [20] . Neither of these two coding schemes considers the off-target situation with bulges.…”

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.…”