Author: Luke Hoberecht
Date: July 24, 2022
The crisprViz
package enables the graphical interpretation of
GuideSet
objects from the
crisprDesign package by
plotting guide RNA (gRNA) cutting locations against their target gene or
other genomic region.
This vignette walks through several use cases that demonstrate the range
of and how to use plotting functions in the crisprViz
package. This
vignette also uses our core gRNA design package
crisprDesignto manipulate
GuideSet
objects in conjunction with plotting in the process of gRNA
design.
Visit our crisprVerse tutorial page to learn more about how to design gRNAs for different applications.
This package is supported for macOS, Linux and Windows machines. Packages were developed and tested on R version 4.2.1
crisprViz
can be installed from Bioconductor using the following
commands in a fresh R session:
install.packages("BiocManager")
BiocManager::install("crisprViz")
All examples in this vignette will use human genome assembly hg38
from
the BSgenome.Hsapiens.UCSC.hg38
package and gene model coordinates
from Ensembl release 104. We begin by loading the necessary packages.
library(BSgenome.Hsapiens.UCSC.hg38)
library(crisprDesign)
library(crisprViz)
Suppose we want to design the four best gRNAs using the SpCas9 CRISPR nuclease to knockout the human KRAS gene. To have the greatest impact on gene function we want to prioritize gRNAs that have greater isoform coverage, and target closer to the 5’ end of the CDS.
Let’s load a precomputed GuideSet
object containing all possible gRNAs
targeting the the CDS of KRAS, and a GRangesList
object describing the
gene model for KRAS.
data("krasGuideSet", package="crisprViz")
data("krasGeneModel", package="crisprViz")
length(krasGuideSet) # number of candidate gRNAs
## [1] 52
For how to design such gRNAs, see the crisprDesign
package. Before we
plot all of our candidate gRNAs, let’s first generate a simple plot with
a few gRNAs to familiarize ourselves with some plot components and
options.
plotGuideSet(krasGuideSet[1:4],
geneModel=krasGeneModel,
targetGene="KRAS")
There are a few things to note here.
- The ideogram track and genome axis track are at the top of our plot and give us coordinate information.
- Our
targetGene
KRAS is plotted next, using coordinates from the provided gene modelkrasGeneModel
, followed by our spacer subset. The name of each track is given on the left. - The strand information for each track is included in the label:
<-
for reverse strand and->
for forward strand. - While we can identify which exon each spacer targets (which may be sufficient), the plot window is too large to provide further information.
- The plot only shows the 3’ end of KRAS, rather than the entire gene.
This last point is important: the default plot window is set by the
spacers’ ranges in the input GuideSet
object. We can manually adjust
this window by using the from
, to
, extend.left
, and extend.right
arguments. Here is the same plot adjusted to show the whole KRAS gene,
which also reveals an additional isoform that is not targeted by any
spacer in this example subset.
from <- min(start(krasGeneModel$transcripts))
to <- max(end(krasGeneModel$transcripts))
plotGuideSet(krasGuideSet[1:4],
geneModel=krasGeneModel,
targetGene="KRAS",
from=from,
to=to,
extend.left=1000,
extend.right=1000)
As calculated above, there are a total of 52 candidate gRNAs targeting
the CDS of KRAS. Including all of them could crowd the plot space,
making it difficult to interpret. To alleviate this we can hide the gRNA
labels by setting the showGuideLabels
argument to FALSE
.
plotGuideSet(krasGuideSet,
geneModel=krasGeneModel,
targetGene="KRAS",
showGuideLabels=FALSE,
from=from,
to=to,
extend.left=1000,
extend.right=1000)
At the gene level, the plot window is too large to discern details for each spacer target. However, we can see five distinct clusters of spacer target locations that cover the CDS of KRAS. The spacers in the 5’-most cluster (on the reverse strand) target the only coding region of the gene that is expressed by all isoforms, making it an ideal target for our scenario.
We can see which gRNAs target this region by returning showGuideLabels
to its default value of TRUE
, and by adjusting the plot window to
focus on our exon of interest.
# new window range around target exon
targetExon <- queryTxObject(krasGeneModel,
featureType="cds",
queryColumn="exon_id",
queryValue="ENSE00000936617")
targetExon <- unique(targetExon)
from <- start(targetExon)
to <- end(targetExon)
plotGuideSet(krasGuideSet,
geneModel=krasGeneModel,
targetGene="KRAS",
from=from,
to=to,
extend.left=20,
extend.right=20)
At this resolution we can get a much better idea of spacer location and orientation. In particular, the PAM sequence is visible as a narrow box on the 3’ side of our protospacer sequences. We can also distinctly see which spacer targets overlap each other–it may be best to avoid pairing such spacers in some applications lest they sterically interfere with each other.
If we have many gRNA targets in a smaller window and are not concerned
with overlaps, we can configure the plot to only show the pam_site
,
rather than the entire protospacer and PAM sequence, by setting
pamSiteOnly
to TRUE
.
plotGuideSet(krasGuideSet,
geneModel=krasGeneModel,
targetGene="KRAS",
from=from,
to=to,
extend.left=20,
extend.right=20,
pamSiteOnly=TRUE)
Let’s filter our GuideSet
by the spacer names in the plot then pass an
on-target score column in our GuideSet
to onTargetScores
to color
the spacers according to that score, with darker blue colors indicating
higher scores. Note that for this plot we need not provide values for
from
and to
, as the plot window adjusts to our filtered GuideSet
.
selectedGuides <- c("spacer_80", "spacer_84", "spacer_88", "spacer_92",
"spacer_96", "spacer_100", "spacer_104", "spacer_108",
"spacer_112")
candidateGuides <- krasGuideSet[selectedGuides]
plotGuideSet(candidateGuides,
geneModel=krasGeneModel,
targetGene="KRAS",
onTargetScore="score_deephf")
For a given CRISPR application, the target region may consist of only several base pairs rather than an exon or entire gene CDS. In these instances it may be important to know exactly where the gRNAs target, and plots of gRNAs must be at a resolution capable of distinguishing individual bases. This is often the case for CRISPR base editor (CRISPRbe) applications, as the editing window for each gRNA is narrow and the results are specific to each target sequence.
In this example, we will zoom in on a few gRNAs targeting the 5’ end of
the human GPR21 gene. We want our plot to include genomic sequence
information so we will set the bsgenome
argument to the same
BSgenome
object we used to create our GuideSet
.
First, we load the precomputed GuideSet
and gene model objects for
GPR21,
data("gpr21GuideSet", package="crisprViz")
data("gpr21GeneModel", package="crisprViz")
and then plot the gRNAs.
plotGuideSet(gpr21GuideSet,
geneModel=gpr21GeneModel,
targetGene="GPR21",
bsgenome=BSgenome.Hsapiens.UCSC.hg38,
margin=0.3)
The genomic sequence is given at the bottom of the plot as color-coded boxes. The color scheme for describing the nucleotides is given in the biovizBase package. If the plot has sufficient space, it will display nucleotide symbols rather than boxes. We can accomplish this by plotting a narrower range or by increasing the width of our plot space (see “Setting plot size” section).
The plot above was generated with a plot space width of 6 inches; here’s the same plot after we increase the width to 10 inches:
# increase plot width from 6" to 10"
plotGuideSet(gpr21GuideSet,
geneModel=gpr21GeneModel,
targetGene="GPR21",
bsgenome=BSgenome.Hsapiens.UCSC.hg38,
margin=0.3)
In this scenario we want to increase expression of the human MMP7 gene via CRISPR activation (CRISPRa). We will use the SpCas9 CRISPR nuclease.
data("mmp7GuideSet", package="crisprViz")
data("mmp7GeneModel", package="crisprViz")
The GuideSet
contains candidate gRNAs in the 2kb window immediately
upstream of the TSS of MMP7. We will also use a GRanges
object
containing repeat elements in this region:
data("repeats", package="crisprViz")
Let’s begin by plotting our GuideSet
, and adding a track of repeat
elements using the annotations
argument. Our guideSet
also contains
SNP annotation, which we would also prefer our gRNAs to not overlap. To
include a SNP annotation track, we will set includeSNPTrack=TRUE
(default).
from <- min(start(mmp7GuideSet))
to <- max(end(mmp7GuideSet))
plotGuideSet(mmp7GuideSet,
geneModel=mmp7GeneModel,
targetGene="MMP7",
guideStacking="dense",
annotations=list(Repeats=repeats),
pamSiteOnly=TRUE,
from=from,
to=to,
extend.left=600,
extend.right=100,
includeSNPTrack=TRUE)
Some of our candidate gRNAs target repeat elements and likely target a large number of loci in the genome, potentially causing unintended effects, or overlap with SNPs, which can reduce its activity. Let’s remove these gRNAs and regenerate the plot.
filteredGuideSet <- crisprDesign::removeRepeats(mmp7GuideSet,
gr.repeats=repeats)
filteredGuideSet <- filteredGuideSet[!filteredGuideSet$hasSNP]
plotGuideSet(filteredGuideSet,
geneModel=mmp7GeneModel,
targetGene="MMP7",
guideStacking="dense",
annotations=list(Repeats=repeats),
pamSiteOnly=TRUE,
from=from,
to=to,
extend.left=600,
extend.right=100,
includeSNPTrack=TRUE)
Note how removing gRNAs that overlap SNPs from our GuideSet
also
removed the SNP track. To prevent plotting an empty track,
plotGuideSet
will only include a SNPs track if at least one gRNA
includes SNP annotation (i.e. overlaps a SNP).
Conversely, there are specific genomic regions that would be beneficial
to target, such as CAGE peaks and DNase I Hypersensitivity tracks. We
show in the inst\scripts
folder how to obtain such data from the
Bioconductor package AnnotationHub
, but for the sake of time, we have
precomputed those objects and they can be loaded from the crisprViz
package directly:
data("cage", package="crisprViz")
data("dnase", package="crisprViz")
We now plot gRNAs alongside with those two tracks:
plotGuideSet(filteredGuideSet,
geneModel=mmp7GeneModel,
targetGene="MMP7",
guideStacking="dense",
annotations=list(CAGE=cage, DNase=dnase),
pamSiteOnly=TRUE,
from=from,
to=to,
extend.left=600,
extend.right=100)
Let’s filter our GuideSet
for guides overlapping the plotted DNase
site then regenerate the plot.
# filter GuideSet for gRNAs overlapping DNase track
overlaps <- findOverlaps(filteredGuideSet, dnase, ignore.strand=TRUE)
finalGuideSet <- filteredGuideSet[queryHits(overlaps)]
plotGuideSet(finalGuideSet,
geneModel=mmp7GeneModel,
targetGene="MMP7",
guideStacking="dense",
annotations=list(CAGE=cage, DNase=dnase),
pamSiteOnly=TRUE,
margin=0.4)
The choice of the CRISPR nuclease can be influenced by the abundance of
PAM sequences recognized by a given nuclease in the target region. For
example, we would expect AT-rich regions to have fewer possible targets
for the SpCas9 nuclease, whose PAM is NGG. In these regions, the CRISPR
nuclease AsCas12a, whose PAM is TTTV, may prove more appropriate. Given
multiple GuideSet
s targeting the same region, we can compare the gRNAs
of each in the same plot using plotMultipleGuideSets
.
Here, we pass our GuideSet
s targeting an exon in the human gene LTN1
in a named list. Note that there are no available options for displaying
guide labels or guide stacking, and only the PAM sites are plotted. We
will also add a track to monitor the percent GC content (using a window
roughly the length of our protospacers). Not surprisingly, this AT-rich
region has fewer targets for SpCas9 compared to AsCas12a. (Note: when
plotting several GuideSets you may need to increase the height of the
plot space in order for the track names to appear on the left side; see
“Setting plot size” below.)
We first load the precomputed GuideSet
objects:
data("cas9GuideSet", package="crisprViz")
data("cas12aGuideSet", package="crisprViz")
data("ltn1GeneModel", package="crisprViz")
plotMultipleGuideSets(list(SpCas9=cas9GuideSet, AsCas12a=cas12aGuideSet),
geneModel=ltn1GeneModel,
targetGene="LTN1",
bsgenome=BSgenome.Hsapiens.UCSC.hg38,
margin=0.2,
gcWindow=10)
Plots with many gene isoforms and/or gRNAs may require more space to render than is allotted by your graphical device’s default settings, resulting in an error. One solution, depending on your graphical device, is offered by the grDevices package.
Here is an example using macOS Quartz device:
grDevices::quartz("Example plot", width=6, height=7)
# plot function
sessionInfo()
## R version 4.2.1 (2022-06-23)
## Platform: x86_64-apple-darwin17.0 (64-bit)
## Running under: macOS Catalina 10.15.7
##
## Matrix products: default
## BLAS: /Library/Frameworks/R.framework/Versions/4.2/Resources/lib/libRblas.0.dylib
## LAPACK: /Library/Frameworks/R.framework/Versions/4.2/Resources/lib/libRlapack.dylib
##
## locale:
## [1] en_US.UTF-8/en_US.UTF-8/en_US.UTF-8/C/en_US.UTF-8/en_US.UTF-8
##
## attached base packages:
## [1] stats4 stats graphics grDevices utils datasets methods
## [8] base
##
## other attached packages:
## [1] crisprViz_0.99.22 crisprDesign_0.99.176
## [3] crisprBase_1.1.8 BSgenome.Hsapiens.UCSC.hg38_1.4.4
## [5] BSgenome_1.65.2 rtracklayer_1.57.0
## [7] Biostrings_2.65.3 XVector_0.37.1
## [9] GenomicRanges_1.49.1 GenomeInfoDb_1.33.7
## [11] IRanges_2.31.2 S4Vectors_0.35.3
## [13] BiocGenerics_0.43.4
##
## loaded via a namespace (and not attached):
## [1] backports_1.4.1 Hmisc_4.7-1
## [3] AnnotationHub_3.5.1 BiocFileCache_2.5.0
## [5] lazyeval_0.2.2 splines_4.2.1
## [7] BiocParallel_1.31.12 ggplot2_3.3.6
## [9] digest_0.6.29 ensembldb_2.21.4
## [11] htmltools_0.5.3 fansi_1.0.3
## [13] checkmate_2.1.0 magrittr_2.0.3
## [15] memoise_2.0.1 cluster_2.1.4
## [17] tzdb_0.3.0 readr_2.1.2
## [19] matrixStats_0.62.0 prettyunits_1.1.1
## [21] jpeg_0.1-9 colorspace_2.0-3
## [23] blob_1.2.3 rappdirs_0.3.3
## [25] crisprScoreData_1.1.3 xfun_0.32
## [27] dplyr_1.0.10 crayon_1.5.1
## [29] RCurl_1.98-1.8 jsonlite_1.8.0
## [31] survival_3.4-0 VariantAnnotation_1.43.3
## [33] glue_1.6.2 gtable_0.3.1
## [35] zlibbioc_1.43.0 DelayedArray_0.23.1
## [37] scales_1.2.1 DBI_1.1.3
## [39] Rcpp_1.0.9 htmlTable_2.4.1
## [41] xtable_1.8-4 progress_1.2.2
## [43] reticulate_1.26 foreign_0.8-82
## [45] bit_4.0.4 Formula_1.2-4
## [47] htmlwidgets_1.5.4 httr_1.4.4
## [49] dir.expiry_1.5.1 RColorBrewer_1.1-3
## [51] ellipsis_0.3.2 pkgconfig_2.0.3
## [53] XML_3.99-0.10 Gviz_1.41.1
## [55] nnet_7.3-17 dbplyr_2.2.1
## [57] deldir_1.0-6 utf8_1.2.2
## [59] tidyselect_1.1.2 rlang_1.0.5
## [61] later_1.3.0 AnnotationDbi_1.59.1
## [63] munsell_0.5.0 BiocVersion_3.16.0
## [65] tools_4.2.1 cachem_1.0.6
## [67] cli_3.4.0 generics_0.1.3
## [69] RSQLite_2.2.16 ExperimentHub_2.5.0
## [71] evaluate_0.16 stringr_1.4.1
## [73] fastmap_1.1.0 yaml_2.3.5
## [75] knitr_1.40 bit64_4.0.5
## [77] purrr_0.3.4 randomForest_4.7-1.1
## [79] AnnotationFilter_1.21.0 KEGGREST_1.37.3
## [81] Rbowtie_1.37.0 mime_0.12
## [83] xml2_1.3.3 biomaRt_2.53.2
## [85] compiler_4.2.1 rstudioapi_0.14
## [87] filelock_1.0.2 curl_4.3.2
## [89] png_0.1-7 interactiveDisplayBase_1.35.0
## [91] tibble_3.1.8 stringi_1.7.8
## [93] crisprScore_1.1.15 highr_0.9
## [95] basilisk.utils_1.9.3 GenomicFeatures_1.49.6
## [97] lattice_0.20-45 ProtGenerics_1.29.0
## [99] Matrix_1.4-1 vctrs_0.4.1
## [101] pillar_1.8.1 lifecycle_1.0.1
## [103] BiocManager_1.30.18 data.table_1.14.2
## [105] bitops_1.0-7 httpuv_1.6.5
## [107] R6_2.5.1 BiocIO_1.7.1
## [109] latticeExtra_0.6-30 promises_1.2.0.1
## [111] gridExtra_2.3 codetools_0.2-18
## [113] dichromat_2.0-0.1 crisprBowtie_1.1.1
## [115] assertthat_0.2.1 SummarizedExperiment_1.27.2
## [117] rjson_0.2.21 GenomicAlignments_1.33.1
## [119] Rsamtools_2.13.4 GenomeInfoDbData_1.2.8
## [121] parallel_4.2.1 hms_1.1.2
## [123] rpart_4.1.16 grid_4.2.1
## [125] basilisk_1.9.6 rmarkdown_2.16
## [127] MatrixGenerics_1.9.1 biovizBase_1.45.0
## [129] Biobase_2.57.1 shiny_1.7.2
## [131] base64enc_0.1-3 interp_1.1-3
## [133] restfulr_0.0.15