Skip to content
/ CRISPRi-seq Public
forked from vincentdebakker/CRISPRi-seq

Design and efficiency evaluation of genome-wide sgRNA libraries for CRISPRi-seq

License

GPL-3.0 license
0 stars 5 forks Branches Tags Activity
Star
Notifications You must be signed in to change notification settings

zyyy97/CRISPRi-seq

BranchesTags
 
 

Repository files navigation

Genome-wide sgRNA library design and evaluation for CRISPRi-seq

Author: Vincent de Bakker
Python code by Afonso Bravo
See also: https://www.veeninglab.com/crispri-seq
Published as a part of:

This repository contains R pipelines to (1) design a de novo CRISPR interference (CRISPRi) single-guide RNA (sgRNA) library and (2) evaluate all potential binding sites of a given sgRNA library on any genome, either provided as GenBank file or NCBI assembly accession number. They were in principle developed for application to prokayotic (specifically bacterial) genomes. For both pipelines there is a version that can be run from within an R environment (we recommend RStudio), and a version that can be called from the command line using the Rscript command. Additionally, the repository contains an example of possible downstream expected efficiency analyses for the latter, as published in Liu et al. (2021). Lastly, it contains a subdirectory with the code to generate the naïve power analysis as published in de Bakker et al. (2022).

General information

The pipelines automatically use multi-core processing, using all available logical processors except one. Input genomes can be either local GenBank files or NCBI assembly accession numbers and multi-chromosome/contig/plasmid genomes are fully supported. sgRNA candidate sequences and binding sites are detected using the CRISPRseek package and match/mismatch vectors between them are defined by the here presented Tool for Identification of Nucleotide-wise, Dual-sequence, Region-delimited identity (TINDRi.py), which is automatically called by the pipelines. Repression activity of an sgRNA at a given site is estimated as the product of within-spacer position-specific penalties of mismatched nucleotides, and referred to as the reprAct score. By using the product, we assume that the observation of Qi et al. (2013) that individual nucleotide mismatch effects on repression activity were independent, extends to the case of >2 and non-adjacent nucleotide mismatches. Although penalties and the reprAct function can be customized by the user, built-in options from the literature are included with penalties from Qi et al. (2013) [E. coli] and Hawkins et al. (2020) [E.coli and B. subtilis].

The pipelines should work on all operating systems, although they have only been tested on Windows 10 and Ubuntu 18.04.5 LTS.

Dependencies

It is necessary to have the following software installed:

  • R version 4.x.x; if the user does not plan to work from the command line, we recommend to install and run from within the RStudio integrated development environment.
  • Python version 3.x; we advise to install the complete Anaconda suite, as this should ensure a proper working python environment.

Installation of packages and modules should be performed automatically the first time a script is invoked (so the script will then also take longer to run that one time). If errors arise nonetheless, please ensure all following dependencies / software packages and modules are installed, manually:

  • R packages CRISPRseek, Biostrings, if using a GenBank file as genome input genbankr, if using an NCBI accession assembly number as genome input biomartr, and if using the *_cmd.R scripts also optparse.
  • Python modules numpy, numba, time, multiprocessing, datetime, os, sys.

de novo sgRNA design

The pipeline in the files sgRNA_library_design.R (invoked within R) and sgRNA_library_design_cmd.R (invoked from command line) can be run to automatically design an sgRNA library for all annotated features in a given genome.

It identifies all possible candidate sgRNAs targeting the non-template strand for all features, as well as all possible sgRNA binding sites on either strand on the whole genome (all contigs/plasmids/chromosomes). Subsequently, it calls the TINDRi.py script, which efficiently matches all pair-wise combinations of sgRNA candidates and binding sites. It detects all relevant sites for each sgRNA based on a user-defined maximum cumulative allowed number of nucleotide mismatches per sub-spacer region, and generates a matrix scoring the mismatches for each relevant sgRNA-site combination. Using that data, the maximum mismatch-based off-target reprAct estimate is computed for each sgRNA, per feature. Finally, the pipeline first selects for each feature the candidate sgRNAs with the lowest expected maximum off-target repression activity (within some user-defined error range of the minimum) and secondly, of that subset, the one sgRNA with the smallest distance to the start codon. Provided that multiple candidate sgRNAs are available, the user may request to design any desired number of sgRNAs per feature.

The pipeline can also exclude or avoid design of sgRNAs with (1) user-defined bad seed sequences, (2) certain motifs in the spacer + primer overhang (in given order of importance, since it might be more crucial to avoid certain restriction enzyme cut sites than e.g. tandem TTTTT motifs) or (3) non-standard bases (i.e., non-ATGC, as for instance some genome sequences contain N's). If desired, the pipeline can output the full list of candidate sgRNAs for all features plus their scores and additional details as well, so that the user may make their own selection. In addition, corresponding forward and reverse oligo's are designed for each selected sgRNA for library construction.

Command line version

Usage

Run:

Rscript sgRNA_library_design_cmd.R --help

to get the user manual and an outline of all flag parameters than can be used. They are all either the same as, or slight adaptations of, the input parameters of sgRNA_library_design.R as described below in more detail. Only the first three input parameters are absolutely required. In its simplest form, the pipeline can thus be run as such:

Rscript sgRNA_library_design_cmd.R -g genome.gb -o ~/output/ -t ~/wd/

where genome.gb is a GenBank file of the genome for which the library is to be designed, ~/output/ is the local directory to write the output files to and ~/wd/ is the local working directory in which TINDRi.py can be found.

Depending on the genome size, number of annotated features, number of allowed mismatches in binding site identification, requested output files and computer specifications (such as number of available logical processors), computation times will likely be in the order of 10-120 minutes.

Output

The pipeline can return multiple files, depending on user input parameters:

  • A .csv file with (an) optimal sgRNA(s) per annotated feature.
  • A .csv file with all candidate sgRNAs for all annotated features.
  • A .csv file with all annotated target features for which no sgRNA could be designed (due to a lack of PAM sequences on the non-template strand).
  • A .txt file with statistics on the run and used input parameters.
  • A .fasta file with all annotated target feature sequences for which sgRNAs were designed.
  • A .fasta file with all candidate sgRNA spacer sequences.
  • A .fasta file with all sgRNA binding site sequences on the genome.
  • A .csv file with all candidate sgRNA - binding site combinations and scores.
  • A .csv file with all candidate sgRNA IDs and sequences, used as input for TINDRi.py.
  • A .csv file with all binding site IDs and sequences, used as input for TINDRi.py.
  • A .csv file with the full mismatch matrix for all candidate sgRNA - binding site combinations, as generated by TINDRi.py. Users will in general be after either or both of the first two files. The third and fourth file are produced by default.

Meaning of variables in sgRNA design output files

Variable Description
target Locus_tag key of the feature the sgRNA was designed to target.
target_name Name of the feature the sgRNA was designed to target.
sgRNA_name Unique name of the sgRNA.
sgRNA_seq Spacer sequence of the sgRNA (PAM-distal to -proximal).
sgRNA_PAM PAM sequence of the on-target binding site of the sgRNA.
sgRNA_GC GC content of the sgRNA spacer, as ratio.
targetlength_bp Length of the target feature in base pairs.
dist2SC_bp Distance of the on-target sgRNA binding site to the target feature start codon in base pairs.
dist2SC_rel Relative distance of the on-target sgRNA binding site to the target feature start codon on the [0,1] interval.
maxOffreprAct Maximum estimated repression activity detected, excluding the on-target binding site the sgRNA was designed for.
all_targets Locus_tags keys of all target features on whose non-template strand the sgRNA has a perfect, zero-mismatch binding site.
all_target_names Names of all target features on whose non-template strand the sgRNA has a perfect, zero-mismatch binding site.
all_chroms Chromosome/plasmid/contig ID of the zero-mismatch binding sites.
all_strands Strand of the zero-mismatch binding sites.
all_ranges Chromosomal locations of the zero-mismatch binding sites, counting from the chromosome start.
oligoForward Forward oligo (primer) to be ordered to construct the annealed sgRNA library pool.
oligoReverse Reverse oligo (primer) to be ordered to construct the annealed sgRNA library pool.
cut_site If any, name of input cut_site sequences detected in the sgRNA spacer; otherwise NA.
nonstandard_nt If any, non-standard nucleotide (non-ATGC) detected in sgRNA spacer; otherwise NA.
bad_seed If any, name of input bad_seed sequences detected in the sgRNA spacer; otherwise NA.
maxOff_mismatches Comma-separated vector of same length as the sgRNA spacer (PAM-distal to -proximal), indicating for each nucleotide whether (1) or not (0) there was a mismatch between that nucleotide in the spacer and the corresponding nucleotide in the off-target binding site with the highest estimated maximum off-target repression activity (as given by maxOffreprAct).
maxOff_seq Sequence (PAM-distal to -proximal) of the off-target binding site with the highest estimated maximum off-target repression activity (as given by maxOffreprAct).

Within-R version

Usage

Open the script sgRNA_library_design.R (we recommend in the RStudio integrated development environment), adjust the input parameters in the first code section #### 1. Settings #### and then run the whole script with those input parameters. In its simplest form, only input_genome, outdir and TINDRidir need to be specified (under ## REQUIRED ##). Depending on the genome size, number of annotated features, number of allowed mismatches in binding site identification, requested output files and computer specifications (such as number of available logical processors), computation times will likely be in the order of 10-120 minutes.

Input

Parameter Description
input_genome required; character. Either the full path to a GenBank file (extension .gb, .gbk, .gbf, or .gbff), or an NCBI assembly accession number (starting with either GCA_ [GenBank] or GCF_ [RefSeq], e.g. GCA_003003495.1).
outdir required character. Path to desired output directory. Should end in a slash '/'.
TINDRidir required character. Path to directory in which TINDRi.py is stored. Should end in a slash '/'.
path_ncbi_downloads character. Path to directory where downloaded files from NCBI should be stored. Only required if input_genome is a GenBank file. Default: NA.
n_sgRNA integer. Desired number of sgRNAs to be designed per annotated feature. Default: 1.
regions integer vector. Indicates the sub-spacer regions for which to define max_mismatch_cum. The sum of these integers must equal the sgRNA spacer length. Default: c(7, 2, 11).
max_mismatch_cum integer vector. Indicates the maximum cumulative number of mismatches allowed over regions for a binding site to be considered relevant. Should be of the same length as regions. For example, max_mismatch_cum of c(1, 2, 11) combined with regions of c(7, 2, 11) allows for 1 mismatch in the first 7 nucleotides (PAM-proximal to -distal), 2 mismatches in the first 7+2=9 nucleotides and 11 mismatches in the first 7+2+11=20 nucleotides (which is the whole sgRNA spacer). Binding sites that do not fulfill these conditions for a given sgRNA are not regarded as relevant and therefore excluded (in the example, a site with e.g. 2 mismatches in the first 7 nucleotides with the spacer would not be considered relevant for that sgRNA, as its repression activity is likely too low at that site). Default: c(1, 2, 11).
reprAct_penalties character or numeric vector. Indicates within-spacer position-dependent penalties for reprAct estimation. For using built-in penalties, must be one of "HawkinsEcoMean", "HawkinsBsuMean", "HawkinsEcoMedian", "HawkinsBsuMedian", "Qi", "QiMean" or "QiMedian". These vectors are the mean, median or raw 20-nt spacer relative repression activity values found by Hawkins et al. (2020) (means and medians per nucleotide) and Qi et al. (2013) (means and medians per region as defined in their paper) for E. coli (Eco) and B. subtilis (Bsu). For custom penalties, please specify a numeric vector (PAM-proximal to -distal) of length sum(regions) (i.e., sgRNA spacer length); if pen_func is prod (recommended), the integers should be on the [0,1] interval. Default: "HawkinsBsuMedian" (PAM-proximal to -distal: 0.35, 0.46, 0.5, 0.53, 0.57, 0.53, 0.57, 0.67, 0.83, 0.9, 0.91, 0.94, 0.95, 0.92, 0.95, 0.84, 0.94, 0.94, 0.96, 0.99).
pen_func character or function. Defines how to process reprAct_penalties of sgRNA-site mismatches to produce reprAct estimates; strong advice not to change. Default: prod (product).
errorRange_maxOffreprAct numeric. The error allowed on maximum off-target repression activity estimates. Candidate sgRNAs with an estimated repression activity within this range from the minimum found off-target repression activity all pass the specificity filter in selecting optimal sgRNAs per feature. This is basically the parameter that weighs the importance of selecting sgRNAs based on estimated off-target activity versus distance of the binding site to the start codon. Setting it to 0 means maximum importance is placed on minimizing off-target activity, so no sgRNA selection based on distance to start codon. Default: 0.4 (40%, as this, in combination with reprAct_penalties set to "HawkinsBsuMedian", reproduced >85% of our manually designed sgRNA library of which we know it works well).
avoidNonBaseNT logical. Whether or not to avoid designing sgRNAs including non-standard bases (so any other than ATGC). Relevant, as genome sequences sometimes include N's, which the user might want to avoid. sgRNAs with non-standard bases are also flagged in the output. Default: TRUE.
bad_seeds character vector. Specifies bad seed sequences to exclude, avoid or flag. Default: c("ACCCA", "TGGAA"), based on Cui et al. (2018).
bad_seed_rule character. Indicates how to treat sgRNAs with bad seed sequences given by bad_seeds. One of "exclude", "avoid" or "ignore", which respectively exclude candidate sgRNAs with these sequences from being designed altogether, avoid their selection if any other candidate sgRNA is available, or just flag sgRNAs with bad seeds in the output. Default: "ignore".
cut_sites named character vector. Indicates in decreasing order of importance motifs to detect in sgRNA spacer + primer sequences (oligoForwardOverhang & oligoReverseOverhang). Could be e.g. specific restriction enzyme cut sites, or tandem T nucleotides to be avoided in design. Reverse-complement sequences are automatically detected too. Default: c(BsmBI = "CGTCTC", tandemT = "TTTTT").
cut_site_rule character. Indicates how to process sgRNAs with specific motifs indicated by cut_sites. One of "exclude", "avoid" or "ignore", which respectively exclude, avoid or just flag sgRNAs with these sequences during design. If "avoid", design of sgRNAs with these motifs is avoided in the order of inputted cut_sites in the presence of alternatives. For example, if for a given feature only sgRNA candidates with BsmBI cut sites and tandem TTTTT motifs exist, the pipeline will prefer to select candidates with the tandem Ts, as per order of default cut_sites input. Default: "avoid".
oligoForwardOverhang character. Post-annealing 5'-3' overhang for the forward oligo to be compatible with the digested sgRNA backbone vector. Default: TATA (BsmBI).
oligoReverseOverhang character. Post-annealing 5'-3' overhang for the reverse oligo to be compatible with the digested sgRNA backbone vector. Default: AAAC (BsmBI).
PAM character. Proto-spacer Adjacent Motif sequence. Default: "NGG" (S. pyogenes dCas9).
filter_out_duplicates logical. Whether or not to filter out duplicate sgRNA spacer sequences in the output list of optimized sgRNAs. Equal sgRNAs can be designed for different annotated features, for instance in the case of gene copies. If set to FALSE, Each designed sgRNA will get its own row in the output, thus appearing multiple times if the same spacer sequence was designed multiple times. Default: TRUE.
output_optimized_list logical. Whether or not to output the .csv file with the list of designed optimal sgRNAs per annotated feature. Note that setting both this parameter and output_all_candidates to FALSE is nonsensical. Default: TRUE.
output_all_candidates logical. Whether or not to output the .csv file with the full list of candidate sgRNAs for all annotated features. Useful if the user wants to explore alternative sgRNAs to be used or to make their own selection from the full list. Default: TRUE.
output_target_fasta logical. Whether or not to output a .fasta file containing the sequences of all found annotated genetic features for which sgRNAs are designed. Default: FALSE.
output_sgRNAs_fasta logical. Whether or not to output a .fasta file containing the sequences of all candidate identified sgRNAs. Default: FALSE.
output_sites_fasta logical. Whether or not to output a .fasta file containing the sequences of all identified sgRNA binding sites on the genome. Default: FALSE.
output_full_list logical. Whether or not to output the complete list of candidate sgRNA - binding site combinations. Note that if set to TRUE, this may take considerable computation time and resources and generate a large .csv file. Not required, recommended to leave set to FALSE. Default: FALSE.
detect_offtarget_genes_full logical. If output_full_list is set to TRUE, whether or not to also detect for each sgRNA-site combination if the site falls within an annotated feature on its non-template strand. Note that if set to TRUE, will add again much more computation time. Not required, recommended to leave set to FALSE. Default: FALSE.
keep_TINDRi_input_sgRNAs logical. Whether or not to keep the intermittently generated .csv file containing all candidate sgRNA IDs and sequences (PAM-proximal to -distal), used as input for TINDRi.py. Default: FALSE.
keep_TINDRi_input_sites logical. Whether or not to keep the intermittently generated .csv file containing all binding site IDs and sequences (PAM-proximal to -distal), used as input for TINDRi.py. Default: FALSE.
keep_TINDRi_matches logical. Whether or not to keep the intermittently generated .csv file containing the full TINDRi.py mismatch output matrix. Note this file is generally large. Default: FALSE.
path_python character. String indicating full path to Python3 installation (e.g. ".../anaconda3/python.exe" on Windows systems). Specify only if the Python3 installation is not automatically detected by the pipeline (it will issue an error message indicating this). Default: NULL (automatic detection).
feature_type character. Feature key tag for each of whose unique identifiers the sgRNAs should be designed. Experimental pipeline feature, in principle do not alter. Default: "locus_tag".

Output

Same as described above for the command line version of the pipeline.

sgRNA binding site evaluation

The pipeline in the files sgRNA_library_evaluation.R (invoked within R) and sgRNA_library_evaluation_cmd.R (invoked on command line) identifies and scores all binding sites up to a specified cumulative region-delimited number of mismatches in a given genome for each sgRNA of a given library.

Much like in the de novo design pipeline, TINDRi.py is called to compare relevant input sgRNA sequences and potential genome-wide binding sites. These results are processed into a repression activity estimate, which is reported alongside other scores.

Command line version

Usage

Run:

Rscript sgRNA_library_evaluation_cmd.R --help

to get the user manual and an outline of all flag parameters than can be used. They are all either the same as, or slight adaptations of, the input parameters of sgRNA_library_evaluation.R as described below in more detail. Only the first four input parameters are absolutely required. In its simplest form, the pipeline can thus be run as such:

Rscript sgRNA_library_evaluation_cmd.R -g genome.gb -s sgRNA-library_spn-D39V.csv -o ~/output/ -t ~/wd/

where genome.gb is a GenBank file of the genome for the library is to be evaluated, sgRNA-library_spn-D39V.csv is a .csv file with the names and spacer sequences (PAM-distal to -proximal, i.e., 5'-3') of the sgRNA library to be evaluated (here our readily available S. pneumoniae D39V library, file available in this repository), ~/output/ is the local directory to write the output files to and ~/wd/ is the local working directory in which TINDRi.py can be found.

Depending on the genome size, number of annotated features, number of allowed mismatches in binding site identification, requested output files and computer specifications (such as number of available logical processors), computation times will likely be in the order of minutes.

Output

The pipeline can return multiple files, depending on user input parameters:

  • A .csv file summarizing per sgRNA the binding sites with first- and second-highest repression activity.
  • A .csv file with all sgRNA - binding site combinations and scores.
  • A .fasta file with all sgRNA spacer sequences.
  • A .fasta file with all sgRNA binding site sequences on the genome.
  • A .csv file with all sgRNA IDs and sequences, used as input for TINDRi.py.
  • A .csv file with all binding site IDs and sequences, used as input for TINDRi.py.
  • A .csv file with the full mismatch matrix for all sgRNA - binding site combinations, as generated by TINDRi.py. Users will in general be after either or both of the first two files.

Meaning of variables in sgRNA evaluation output files

Not all variables exist in all output files.

Variable Description
sgRNA_ID Unique sgRNA identifier, used within the pipeline.
mismatches Comma-separated string representing mismatches (1) and matches (0) per within-spacer position (PAM-proximal to -distal)
site_ID Unique binding site identifier, used within the pipeline.
sgRNA_name Unique sgRNA name.
n_mismatch Total number of mismatches between given sgRNA spacer and binding site.
reprAct Estimated mismatch-based repression activity of a given sgRNA on a given site.
chrom Name of the chromosome on which the binding site is located.
strand Strand of the chromosome on which the binding site is located.
start Start base pair of the binding site location on the given chromosome.
end End base pair of the binding site location on the given chromosome.
sgRNA_seq Sequence of the sgRNA spacer (PAM-distal to -proximal).
site_seq Sequence of the binding site (PAM-distal to -proximal).
site_PAM Proto-spacer Adjacent Motif (PAM) sequence of the binding site.
sgRNA_GC GC content of the sgRNA spacer, as a ratio.
all_targets Locus_tag keys of all features within whose boundaries the binding site is located on the non-template strand, if any; otherwise NA.
all_target_names If available, names of all features within whose boundaries the binding site is located on the non-template strand. Otherwise locus_tag keys, if any; otherwise NA.
all_chroms Chromosome names on which binding site is located.
all_strands DNA strands on which binding site is located.
all_ranges Genomic locations in bp (formatted as start..end) of the binding site.
bad_seed If any, name of input --bad_seeds sequences detected in the sgRNA spacer; otherwise NA.
cut_site If any, name of input --cut_sites sequences detected in the sgRNA spacer; otherwise NA.
reprAct_1st / reprAct_2nd Estimated mismatch-based repression activity of the sgRNA at the binding sites with first- and second-highest estimated repression activities.
nmismatch_1st / nmismatch_2nd Total number of mismatches between the sgRNA spacer and the sequences of the binding sites with first- and second-highest estimated repression activities.
targets_1st / targets_2nd Respectively for the binding site with the first- (reprAct_1st) and second-highest (reprAct_2nd) estimated repression activity, the locus_tag keys of features within whose boundaries the binding site is located on the non-template strand, if any; otherwise NA.
target_names_1st / target_names_2nd If available, names of features corresponding to targets_1st and targets_2nd, otherwise locus_tag keys. If none, NA.
chroms_1st / chroms_2nd Chromosome name on which the binding sites with first- and second-highest estimated repression activities are located.
strands_1st / strands_2nd DNA strands on which the binding sites with first- and second-highest estimated repression activities are located.
ranges_1st / ranges_2nd Genomic location in bp (formatted as start..end) of the binding sites with first- and second-highest estimated repression activities.
seq_1st / seq_2nd Sequences of the binding sites with first- and second-highest estimated repression activities.
PAM_1st / PAM_2nd PAM sequences of the binding sites with first- and second-highest estimated repression activities.

Within-R version

Usage

Open the script sgRNA_library_evaluation.R (we recommend in the RStudio integrated development environment), adjust the input parameters in the first code section #### 1. Settings #### and then run the whole script with those input parameters. In its simplest form, only input_genome, sgRNA_file, outdir and TINDRidir need to be specified (under ## REQUIRED ##). Depending on the genome size, number of annotated features, number of allowed mismatches in binding site identification, requested output files and computer specifications (such as the number of available logical processors for multi-core processing), computation times may be in the order of minutes.

Input

Parameter Description
input_genome required; character. Either the full path to a GenBank file (extension .gb, .gbk, .gbf, or .gbff), or an NCBI assembly accession number (starting with either GCA_ [GenBank] or GCF_ [RefSeq], e.g. GCA_003003495.1).
sgRNA_file required; character. Full path to a header-less .csv file with two comma-separated columns, the first containing sgRNA-unique IDs (e.g. sgRNA0001, sgRNA0002, etc.) and the second containing the sgRNA spacer sequences (PAM-distal to -proximal, i.e., 5 '-3'). As sgRNA-library_spn-D39V.csv in this repository.
outdir required character. Path to desired output directory. Should end in a slash '/'.
TINDRidir required character. Path to directory in which TINDRi.py is stored. Should end in a slash '/'.
path_ncbi_downloads character. Path to directory where downloaded files from NCBI should be stored. Only required if input_genome is a GenBank file. Default: NA.
regions integer vector. Indicates the sub-spacer regions for which to define max_mismatch_cum. The sum of these integers must equal the sgRNA spacer length. Default: c(7, 2, 11).
max_mismatch_cum integer vector. Indicates the maximum cumulative number of mismatches allowed over regions for a binding site to be considered relevant. Should be of the same length as regions. For example, max_mismatch_cum of c(1, 2, 11) combined with regions of c(7, 2, 11) allows for 1 mismatch in the first 7 nucleotides (PAM-proximal to -distal), 2 mismatches in the first 7+2=9 nucleotides and 11 mismatches in the first 7+2+11=20 nucleotides (which is the whole sgRNA spacer). Binding sites that do not fulfill these conditions for a given sgRNA are not regarded as relevant and therefore excluded (in the example, a site with e.g. 2 mismatches in the first 7 nucleotides with the spacer would not be considered relevant for that sgRNA, as its repression activity is likely too low at that site). Default: c(1, 2, 11).
reprAct_penalties character or numeric vector. Indicates within-spacer position-dependent penalties for reprAct estimation. For using built-in penalties, must be one of "HawkinsEcoMean", "HawkinsBsuMean", "HawkinsEcoMedian", "HawkinsBsuMedian", "Qi", "QiMean" or "QiMedian". These vectors are the mean, median or raw 20-nt spacer relative repression activity values found by Hawkins et al. (2020) (means and medians per nucleotide) and Qi et al. (2013) (means and medians per region as defined in their paper) for E. coli (Eco) and B. subtilis (Bsu). For custom penalties, please specify a numeric vector (PAM-proximal to -distal) of length sum(regions) (i.e., sgRNA spacer length); if pen_func is prod (recommended), the integers should be on the [0,1] interval. Default: "HawkinsBsuMedian" (PAM-proximal to -distal: 0.35, 0.46, 0.5, 0.53, 0.57, 0.53, 0.57, 0.67, 0.83, 0.9, 0.91, 0.94, 0.95, 0.92, 0.95, 0.84, 0.94, 0.94, 0.96, 0.99).
pen_func character or function. Defines how to process reprAct_penalties of sgRNA-site mismatches to produce reprAct estimates; strong advice not to change. Default: prod (product).
bad_seeds character vector. Specifies bad seed sequences to flag. Default: c("ACCCA", "TGGAA"), based on Cui et al. (2018).
cut_sites named character vector. Indicates the motifs to detect in sgRNA spacer + primer sequences (oligoForwardOverhang & oligoReverseOverhang). Could be e.g. specific restriction enzyme cut sites, or tandem T nucleotides one wishes to avoid. Reverse-complement sequences are automatically detected too. Default: c(BsmBI = "CGTCTC", tandemT = "TTTTT").
oligoForwardOverhang character. Post-annealing 5'-3' overhang for the forward oligo to be compatible with the digested sgRNA backbone vector. Default: TATA (BsmBI).
oligoReverseOverhang character. Post-annealing 5'-3' overhang for the reverse oligo to be compatible with the digested sgRNA backbone vector. Default: AAAC (BsmBI).
PAM character. Proto-spacer Adjacent Motif sequence. Default: "NGG" (S. pyogenes dCas9).
output_exact logical. Whether or not to output the main results summary output table as .csv file, containing per sgRNA bad seed and cut site flags, GC content and details on the binding sites with the first- and second-highest estimated maximum repression activity. Note that setting both this and output_full to FALSE is nonsensical. Default: TRUE.
output_full logical. Whether or not to output a .csv file with the complete list of candidate sgRNA - binding site combinations & detailed information per combination. Note that for very large sgRNA libraries, this may take considerable computation time and resources and generate a large file. Default: FALSE.
output_sgRNAs_fasta logical. Whether or not to output a .fasta file containing the sequences of all sgRNAs. Default: FALSE.
output_sites_fasta logical. Whether or not to output a .fasta file containing the sequences of all identified sgRNA binding sites on the genome. Default: FALSE.
detect_offtarget_genes_full logical. If output_full is set to TRUE, whether or not to also detect for each sgRNA-site combination if the site falls within an annotated feature on its non-template strand. Note that if set to TRUE, this might add computation time, especially for large sgRNA libraries. Default: FALSE.
keep_TINDRi_input_sgRNAs logical. Whether or not to keep the intermittently generated .csv file containing all candidate sgRNA IDs and sequences (PAM-proximal to -distal), used as input for TINDRi.py. Default: FALSE.
keep_TINDRi_input_sites logical. Whether or not to keep the intermittently generated .csv file containing all binding site IDs and sequences (PAM-proximal to -distal), used as input for TINDRi.py. Default: FALSE.
keep_TINDRi_matches logical. Whether or not to keep the intermittently generated .csv file containing the full TINDRi.py mismatch output matrix. Note this file is generally large. Default: FALSE.
path_python character. String indicating full path to Python3 installation (e.g. ".../anaconda3/python.exe" on Windows systems). Specify only if the Python3 installation is not automatically detected by the pipeline (it will issue an error message indicating this). Default: NULL (automatic detection).
feature_type character. GenBank key tag of features to be detected. Experimental pipeline feature, in principle do not alter. Default: "locus_tag".

Output

Same as described above for the command line version of the pipeline.

Downstream efficiency analyses

The file Pneumococcal_sgRNA_library_efficiency_evaluation.pdf contains downstream analyses of the results of a former version of sgRNA_library_evaluation.R for seven pneumococcal genomes, given the sgRNA library sgRNA-library_spn-D39V.csv of Liu et al. (2020). For the binding site results, see https://www.veeninglab.com/crispri-seq. Source code of the pdf: Pneumococcal_sgRNA_library_efficiency_evaluation.Rmd.

References

  • Liu, X., Kimmey, J.M., Matarazzo, L., de Bakker, V., Van Maele, L., Sirard, J.-C., Nizet, V., Veening, J.-W. (2021) Exploration of Bacterial Bottlenecks and Streptococcus pneumoniae pathogenesis by CRISPRi-seq. Cell Host Microbe 29, 107-120.e6. https://doi.org/10.1016/j.chom.2020.10.001.
  • de Bakker, V., Liu, X., Bravo, A.M., Veening, J.-W. (2022) CRISPRi-seq for genome-wide fitness quantification in bacteria. Nature Protocols 17, 252–281. https://doi.org/10.1038/s41596-021-00639-6.
  • Qi, L.S., Larson, M.H., Gilbert, L.A., Doudna, J.A., Weissman, J.S., Arkin, A.P., Lim, W.A. (2013) Repurposing CRISPR as an RNA-guided platform for sequence-specific control of gene expression. Cell 152, 1173–1183. https://doi.org/10.1016/j.cell.2013.02.022
  • Hawkins, J.S., Silvis, M.R., Koo, B-M., Peters, J.M., Osadnik, H., Jost, M., Hearne, C.C., Weissman, J.S., Todor, H., Gross, C.A. (2020) Mismatch-CRISPRi Reveals the Co-varying Expression-Fitness Relationships of Essential Genes in Escherichia coli and Bacillus subtilis. Cell Syst. 11, 523-535.e9. https://doi.org/10.1016/j.cels.2020.09.009
  • Cui, L., Vigouroux, A., Rousset, F., Varet, H., Khanna, V., Bikard, D. (2018) A CRISPRi screen in E. coli reveals sequence-specific toxicity of dCas9. Nat. Commun. 9, 1912. http://dx.doi.org/10.1038/s41467-018-04209-5
  • Zhu, L.J., Holmes, B.R., Aronin, N., Brodsky, M.H. (2014) CRISPRseek: A Bioconductor Package to Identify Target-Specific Guide RNAs for CRISPR-Cas9 Genome-Editing Systems. PLoS ONE 9(9): e108424. https://doi.org/10.1371/journal.pone.0108424
  • Drost, H.-G., Paszkowski, J. (2017) Biomartr: genomic data retrieval with R. Bioinformatics 33(8), 1216–1217. https://doi.org/10.1093/bioinformatics/btw821
  • Trevor L Davis (2020). optparse: Command Line Option Parser. R package version 1.6.6. https://CRAN.R-project.org/package=optparse
  • Gabriel Becker and Michael Lawrence (2020). genbankr: Parsing GenBank files into semantically useful objects. R package version 1.16.0. https://doi.org/doi:10.18129/B9.bioc.genbankr
  • H. Pagès, P. Aboyoun, R. Gentleman and S. DebRoy (2020). Biostrings: Efficient manipulation of biological strings. R package version 2.56.0. https://doi.org/doi:10.18129/B9.bioc.Biostrings

About

Design and efficiency evaluation of genome-wide sgRNA libraries for CRISPRi-seq

Resources

Readme

License

GPL-3.0 license
Activity

Stars

0 stars

Watchers

0 watching

Forks

0 forks
Report repository

Releases

No releases published

Packages

No packages published

Languages

  • R 97.0%
  • Python 3.0%