2 Multicrispr

Multicrispr aism to be a gRNA design solution which is:

complete: defining/transforming targets, finding spacers, analysing on/off-targets
performant: scales towards thousands of targets (data.table backend)
intuitive: a functional programming design and intuitive visualizations
interoperable: acceping and returning a GRanges

For on-target scoring either the Doench2014 or the Doench2016 method can be used (Doench2016 is the current standard, see e.g. Haeussler et al., 2016). For off-target analysis Bowtie (fast) as well as vcountPDict (exact) can be used. The figure below gives an overview of how multicrispr can be used, the subsequent sections below discuss the details.

2.1 Install

Installing multicrispr is straightforward:

# From BioC
install.packages("BiocManager")
BiocManager::install(version='devel')
BiocManager::install("multicrispr")

# From gitlab: 
#url <- 'https://gitlab.gwdg.de/loosolab/software/multicrispr.git'
#remotes::install_git(url, repos = BiocManager::repositories())

Doench et al. (2016) ’s python package azimuth for on-target efficiency prediction using their method can be easily installed and activated using reticulate:

# Install once
  # reticulate::conda_create('azienv', 'python=2.7')
  # reticulate::conda_install('azienv', 'azimuth', pip = TRUE)
  # reticulate::conda_install('azienv', 'scikit-learn==0.17.1', pip = TRUE)
# Then activate
  reticulate::use_condaenv('azienv')

Bowtie-indexed genomes for fast offtarget analysis can be installed using index_genome. For the two genomes used in the examples, mm10 and hg38, the function downloads pre-build indexes from our data server, allowing a quick exploration (set download=FALSE to build index anew):

index_genome(BSgenome.Mmusculus.UCSC.mm10::BSgenome.Mmusculus.UCSC.mm10)
index_genome(BSgenome.Hsapiens.UCSC.hg38::BSgenome.Hsapiens.UCSC.hg38  )

2.2 Define targets

bed_to_granges converts a (0-based) BED coordinate file into a (1-based) GRanges.
An example is loading the 1974 binding sites of the transcription factor SRF:

require(magrittr)
require(multicrispr)
bedfile <- system.file('extdata/SRF.bed', package = 'multicrispr')
targets0 <- bed_to_granges(bedfile, genome = 'mm10')

genes_to_granges and genefile_to_granges convert entrez/ensembl gene identifiers into GRanges using txdb gene models. An example that loads a small subset of the 1974 SRF binding sites:

entrezfile <- system.file('extdata/SRF.entrez', package = 'multicrispr')
txdb <- TxDb.Mmusculus.UCSC.mm10.knownGene::TxDb.Mmusculus.UCSC.mm10.knownGene
invisible(genefile_to_granges(entrezfile, txdb, complement = TRUE))

char_to_granges uses a 1-based coordinate vector to specify the GRanges.
It can be used e.g. to define the first SRF binding site explicitly:

bsgenome <- BSgenome.Mmusculus.UCSC.mm10::BSgenome.Mmusculus.UCSC.mm10
plot_intervals(char_to_granges(c(srf1 = 'chr13:119991554-119991569:+'), bsgenome))

2.3 Transform

As a second step, the original targets may require transformation. The functions up_flank (target upstream flanks), down_flank (downstream flanks), double_flank (double flanks) and extend can be used to update the target ranges as required.

# Up flank
    invisible(up_flank(  targets0, -200, -1))
# Down flank
    invisible(down_flank( targets0, 1, 200))
# Double flank
    invisible(double_flank(targets0, -200, -1, +1, +200))
# Extend
    targets <- extend(targets0, -22, 22, plot = TRUE)

2.4 Find spacers, count off-targets, score on-targets

find_spacers finds N20 spacers (followed by NGG pam sites), counts off-targets, and scores on-targets. It also visualizes the results and returns them as a GRanges object with off-target counts and on-target efficiency.

spacers <- find_spacers(
    targets, bsgenome, complement=FALSE, mismatches=0, subtract_targets = TRUE)

2.5 Return/Write

The results are returned as a GRanges object. They can be written to file with write_ranges
They can be cast into a data.table with gr2dt, as shown below

  str(gr2dt(spacers), vec.len=2)
## Classes 'data.table' and 'data.frame':   6616 obs. of  15 variables:
##  $ seqnames    : Factor w/ 239 levels "chr1","chr2",..: 1 1 1 1 1 ...
##  $ start       : int  4712619 4712620 4712627 4712633 4712634 ...
##  $ end         : int  4712638 4712639 4712646 4712652 4712653 ...
##  $ width       : int  20 20 20 20 20 ...
##  $ strand      : Factor w/ 3 levels "+","-","*": 2 2 2 2 2 ...
##  $ name        : chr  "SRF_MA0083.3" "SRF_MA0083.3" ...
##  $ score       : num  10.5 10.5 ...
##  $ targetname  : chr  "chr1:4712628-4712643:-" "chr1:4712628-4712643:-" ...
##  $ targetstart : int  4712628 4712628 4712628 4712628 4712628 ...
##  $ targetend   : int  4712643 4712643 4712643 4712643 4712643 ...
##  $ crisprname  : chr  "chr1:4712628-4712643:-_5" "chr1:4712628-4712643:-_4" ...
##  $ crisprspacer: chr  "ATATAAGGGCATTGGAAGAA" "AATATAAGGGCATTGGAAGA" ...
##  $ crisprpam   : chr  "GGG" "AGG" ...
##  $ Doench2014  : num  0.16491 0.00721 ...
##  $ names       : chr  "chr1:4712628-4712643:-_5" "chr1:4712628-4712643:-_4" ...
##  - attr(*, ".internal.selfref")=<externalptr>

Crispr/Cas9 gRNA design

2024-05-01

Contents

1 Background