deepCpf1: DeepCpf1 Algorithm for predicting CRISPR-Cpf1 gRNA Efficacy

In LihuaJulieZhu/CRISPRseek: Design of target-specific guide RNAs in CRISPR-Cas9, genome-editing systems

DeepCpf1 Algorithm for predicting CRISPR-Cpf1 gRNA Efficacy

Description

DeepCpf1 algorithm from https://doi.org/10.1038/nbt.4061, which takes in 34 bp target sequences with/without chromatin accessibility information and returns predicted CRISPR-Cpf1 gRNA efficacy for each input sequence.

Usage

deepCpf1(extendedSequence, chrom_acc)

Arguments

extendedSequence	Sequences containing gRNA plus PAM plus flanking sequences. Each sequence should be 34 bp long as specified by http://deepcrispr.info/, i.e., 4bp before the 5' PAM, 4bp PAM, 20bp gRNA, and 6bp after 3' of gRNA.
chrom_acc	Optional binary variable indicating chromatin accessibility information with 1 indicating accessible and 0 not accessible.

Details

Having chromatin accessibility information will aid in the accuracy of the scores, but one can still get accurate scoring with only the 34 bp target sequences.

Value

a numeric vector with prediced CRISPR-Cpf1 gRNA efficacy taking into account chromatin accessibility information if accessibility information is provided

Author(s)

Paul Scemama and Lihua Julie Zhu

References

Kim et al., Deep learning improves prediction of CRISPR–Cpf1 guide RNA activityNat Biotechnol 36, 239–241 (2018). https://doi.org/10.1038/nbt.4061

Examples

library(keras)
library(mltools)
library(dplyr)
library(data.table)

use_implementation("tensorflow")

extendedSequence <- c('GTTATTTGAGCAATGCCACTTAATAAACATGTAA',
 'TGACTTTGAATGGAGTCGTGAGCGCAAGAACGCT',
 'GTTATTTGAGCAATGCCACTTAATAAACATGTAA',
 'TGACTTTGAATGGAGTCGTGAGCGCAAGAACGCT')
chrom_acc <- c(0,1, 0, 1)

if (interactive()) {
 deepCpf1(extendedSequence = extendedSequence, chrom_acc = chrom_acc)
}

LihuaJulieZhu/CRISPRseek documentation built on Feb. 3, 2024, 2:44 p.m.