Deep learning improves prediction of CRISPR–Cpf1 guide RNA activity

Abstract

We present two algorithms to predict the activity of AsCpf1 guide RNAs. Indel frequencies for 15,000 target sequences were used in a deep-learning framework based on a convolutional neural network to train Seq-deepCpf1. We then incorporated chromatin accessibility information to create the better-performing DeepCpf1 algorithm for cell lines for which such information is available and show that both algorithms outperform previous machine learning algorithms on our own and published data sets.

Massively parallel CRISPRi assays reveal concealed thermodynamic determinants of dCas12a binding

“…), suggesting that the specific base compo-sition of the seed region does not have a large impact on DNA binding. This contrasts with in vivo multiplexed DNA cleavage assays for Cas12a variants that do show significant sequence dependence on cleavage activity [15,19,20]. In addition, while SpCas9 binding and cleavage activity has different sequence specificities [12][13][14], we do not observe any significant discrepancies between the binding and cleavage assays performed using catalytically-active Fn-Cas12a nuclease (Fig.…”

“…This is especially important in the context of CRISPR base editors [10,11] because off-target binding, which may not entirely correlate with DNA cleavage [12][13][14], needs to be reduced to a minimum level * Corresponding author to prevent unintended base changes. While several in silico models [15][16][17][18][19][20] have been developed to predict the binding affinity of RNA guided CRISPR-Cas proteins using data from in vitro biochemical assays [21][22][23][24] or in vivo indel frequencies [12][13][14][25][26][27], these approaches only provide empirical interpretations of CRISPR-Cas DNA binding and often fail to yield a conceptual understanding of the underlying factors involved in CRISPR-Cas binding. Furthermore, it can be difficult to extract quantitative binding affinity measurements from in vivo indel frequencies due to the inherent CRISPR-Cas binding inefficiencies associated with cellular physiological factors such as cell type, chromatin state, and delivery method [28][29][30].…”

Improved CRISPR‐Cas12a‐assisted one‐pot DNA editing method enables seamless DNA editing

“…On the other hand, several crRNA design tools have been developed for CRISPR systems to deal with the off-target or low-efficiency issues (Labun, Montague, Gagnon, Thyme, & Valen, 2016;Moreno-Mateos et al, 2015;Xie, Shen, Zhang, Huang, & Zhang, 2014). Recently, a design tool was reported that can enable accurate prediction of AsCas12a crRNA activities in 125 cells (Kim et al, 2018). With further understanding of the guide RNA activities in vitro, the selection of crRNAs for iCOPE could be standardized and optimized.…”

Versatile and robust genome editing with Streptococcus thermophilus CRISPR1-Cas9

“…However, many such enzymes were found inactive in human cells despite being accurately reprogrammed for DNA binding and cleavage in vitro [7][8][9][10] . Nevertheless, the full potential of selected enzymes can be unleashed using machine learning to establish sgRNA design rules 11,12 .…”

“…However, many such enzymes were found inactive in human cells despite being accurately reprogrammed for DNA binding and cleavage in vitro [7][8][9][10] . Nevertheless, the full potential of selected enzymes can be unleashed using machine learning to establish sgRNA design rules 11,12 .Perhaps the most striking example of the value of alternative Cas9 enzymes is the implementation of the type II-A Cas9 from Staphylococcus aureus (SaCas9) for in vivo editing using recombinant adeno-associated virus (rAAV) vectors 7,13, 14 . More recently, Campylobacter jejuni and Neisseria meningitidis Cas9s from the type II-C 15 CRISPRCas systems have been added to this repertoire 16,17 .…”