`guido`

Submodules

Package Contents

Classes

`Genome`
`Guide`
`Locus`

Functions

`load_genome_from_file`(guido_file)	Load a genome from a .guido file. This file is created when a genome is
`is_tool`(name)	Check whether name is on PATH and marked as executable.
`rev_comp`(seq[, rna])
`locus_from_coordinates`(genome, chromosome, start, end)	Create a locus from coordinates. Coordinates are 1-based. If annotation
`locus_from_gene`(genome, gene_name)	Create a locus from gene name. If annotation file is provided, it will
`locus_from_sequence`(sequence[, sequence_name])	Create a locus from sequence.

class guido.Genome(genome_name, genome_file_abspath=None, annotation_file_abspath=None, bowtie_index_abspath=None)[source]

property is_built

Check if genome index is built.

Returns:

Bool: True if genome index is built, False otherwise.

property sequence

build(n_threads=1, save_pickle=True, bowtie_ignore=False, bowtie_path='')[source]

Build genome index.

This method creates index files for the genome and annotation files and saves them in a .guido file in the same directory as the genome Fasta file. This file can be later loaded using guido.genome.load_genome_from_file() without having to re-build the index.

Parameters:

n_threadsint, optional: Number of threads, by default 1
save_picklebool, optional: Pickle the dictionary into .guido file, by default True
bowtie_ignorebool, optional: Ignore building bowtie index. Use if you already have bowtie index built, by default False
bowtie_pathstr, optional: Path to bowtie binary if it’s not in the path, by default “”

guido.load_genome_from_file(guido_file)[source]

Load a genome from a .guido file. This file is created when a genome is built. It contains all the information needed to use the genome. Guido files are saved in the same directory as the genome FASTA file. They are named after the genome name and have the .guido extension. Genome object can be created by using the :method:`build` method.

Parameters:

guido_filestr: Path to the .guido file.

Returns:

Genome: Genome object.

class guido.Guide(sequence, pam_position, pam_len, strand='+', max_flanking_length=75, cut_offset=3, chromosome='seq', start=0)[source]

property location

Returns the location of the guide in the format: chr:start-end.

Returns:

str: String representation of gRNA location

property off_targets_string

Returns a string representation of the off-targets.

The string representatio captures the number of off-targets with certain number of mismatches: n0|n1|n2|n3|n4|n5 (total), where n0 is the number of off-targets with 0 mismatches, n1 is the number of off-targets with 1 mismatch, etc.

For example, if there are 3 off-targets with 0 mismatches, 2 with 1 mismatch, 1 with 2 mismatches, 0 with 3 mismatches, 5 with 4 mismatches and 1 with 5 mismatches the string representation will be “3|2|1|0|5|1 (13)”. In the parenthesis, the total number of off-targets is given.

Returns:

str: String representation of off-targets.

property layers

__repr__()[source]: Return repr(self).

__getattr__(attr)[source]

_create_mmej_oof_string(mmej_patterns)[source]

simulate_end_joining(n_patterns=5, length_weight=20)[source]

Simulate Microhomology-Mediated End Joining (MMEJ) events for the gRNA.

MMEJ scoring is based on the Bae et al. 2014 paper (https://doi.org/10.1038/nmeth.3015)

Parameters:

n_patternsint, optional: Number of top-scoring MMEJ patterns to keep, by default 5
length_weightint, optional: Lengeth weight, by default 20

find_off_targets(genome, **kwargs)[source]

Finds off-targets for the guide. The off-targets are found using Bowtie. Bowtie index for the genome must be built before running this function.

Parameters:

genomeGenome: Genome object with the Bowtie index built

Notes

The off-targets are stored in the off_targets attribute. Based on the off-targets, the following layers are added to the guide:

ot_sum_score: sum of the off-target scores - the lower the better
ot_cfd_score_mean: mean of the CFD scores of the off-targets
ot_cfd_score_max: max CFD scores of the off-targets
ot_cfd_score_sum: sum CFD scores of the off-targets

add_layer(name, layer_data)[source]

_summary_

Parameters:

namestr: _description_
layer_datafloat: _description_

layer(key)[source]

_summary_

Parameters:

key_type_: _description_

Returns:

_type_: _description_

Raises:

ValueError: _description_

add_azimuth_score(model_file='V3_model_nopos.pickle')[source]

Apply Azimuth score to a list of guides.

Azimuth is a machine learning-based predictive modelling of CRISPR/Cas9 guide efficiency. Sometimes its reffered to as Doench 2016 score.

Described in https://doi.org/10.1038/nbt.3437 (Doench et al., 2016)

Returns:

float: Azimuth score.

guido.is_tool(name)[source]: Check whether name is on PATH and marked as executable.

guido.rev_comp(seq, rna=False)[source]

class guido.Locus(sequence, name=None, start=1, end=None, genome=None, annotation=None, **kwargs)[source]

property layers

Layers of the locus.

Returns:

Layers: List of layers

__repr__()[source]: Returns a string representation of the locus object.

to_dict()[source]: Converts the locus object to a dictionary.

guide(ix)[source]

Fetch a guide from the locus by its index or name.

Parameters:

ixstr or int: Index of the gRNA.

Returns:

g: Guide: Guide object representing a gRNA

Examples

>>> import guido
>>> seq = "TTATCATCCACTCTGACGGGTGGTATTGCGCAACTCCACGCCATCAAACATGTTCAGATTATGCAATCGTGAGTATTCGTTGACCACCGCTTGACCTGTGT"
>>> loc = guido.Locus(
...     sequence=seq, name="AgamP4_2R", start=48714554, end=48714654
... )
>>> loc.find_guides()
>>> loc.guide("gRNA-1")
gRNA-1(CGCAATACCACCCGTCAGAGTGG|AgamP4_2R:48714561-48714584|-|)
>>> loc.guide(0)
gRNA-1(CGCAATACCACCCGTCAGAGTGG|AgamP4_2R:48714561-48714584|-|)

_flatten_intervals(intervals)[source]: Flattens overlapping intervals into an union.

_find_guides_in_interval(sequence, start, pam)[source]: Finds guides in interval.

find_guides(pam='NGG', min_flanking_length=0, selected_features='all')[source]

Find gRNAs in the locus.

Parameters:

pamstr, optional: gRNA PAM sequence, by default “NGG”
min_flanking_lengthint, optional: Defines flanking region from the locus where gRNAs are ignored. By default 0, however simulate_end_joining() requires flanking region of 75 bp to simulate MMEJ.
selected_featuresstr, optional: Limit gRNA search on only specified genomic features. Features are defined in the provided genome annotation file. By default {“all”}

Returns:

sorted_guideslist: List of gRNAs sorted by their position in the locus.

Examples

>>> import guido
>>> genome = guido.load_genome_from_file(
...     guido_file="/Users/nkranjc/imperial/ref/new/AgamP4.guido"
... )
>>> loc = guido.locus_from_coordinates(genome, "AgamP4_2R", 48714541, 48714666)
>>> loc.find_guides()
>>> loc.guides
[gRNA-1(AAGTTTATCATCCACTCTGACGG|AgamP4_2R:48714550-48714572|+|),
gRNA-2(CGCAATACCACCCGTCAGAGTGG|AgamP4_2R:48714561-48714583|-|),
gRNA-3(AGTTTATCATCCACTCTGACGGG|AgamP4_2R:48714551-48714573|+|),
gRNA-4(TTATCATCCACTCTGACGGGTGG|AgamP4_2R:48714554-48714576|+|),
gRNA-5(TCTGAACATGTTTGATGGCGTGG|AgamP4_2R:48714589-48714611|-|),
gRNA-6(CATAATCTGAACATGTTTGATGG|AgamP4_2R:48714594-48714616|-|),
gRNA-7(GTTTAACACAGGTCAAGCGGTGG|AgamP4_2R:48714637-48714659|-|),
gRNA-8(TATGTTTAACACAGGTCAAGCGG|AgamP4_2R:48714640-48714662|-|)]

Searching for gRNAs in a specific genomic feature:

>>> loc.find_guides(selected_features="exon")
>>> loc.guides
[gRNA-1(AAGTTTATCATCCACTCTGACGG|AgamP4_2R:48714550-48714572|+|),
gRNA-2(CGCAATACCACCCGTCAGAGTGG|AgamP4_2R:48714561-48714583|-|),
gRNA-3(AGTTTATCATCCACTCTGACGGG|AgamP4_2R:48714551-48714573|+|),
gRNA-4(TTATCATCCACTCTGACGGGTGG|AgamP4_2R:48714554-48714576|+|),
gRNA-5(TCTGAACATGTTTGATGGCGTGG|AgamP4_2R:48714589-48714611|-|),
gRNA-6(CATAATCTGAACATGTTTGATGG|AgamP4_2R:48714594-48714616|-|)]

simulate_end_joining(n_patterns=5, length_weight=20)[source]

Simulate end-joining and find MMEJ deletion patterns for each gRNA.

Microhomology scores are calculated based on proposed scoring model described by Bae et al. 2014.

Parameters:

n_patternsint, optional: Number of top scored MMEJ deletion patterns reported. By default 5.
length_weightint, optional: Length weight parameter used in MMEJ scoring as defined by Bae et al. 2015. By default, 20.

find_off_targets(external_genome=None, **kwargs)[source]

Find off-targets in the genome for each gRNA.

Parameters:

external_genomeGenome, optional: If provided, off-target search is performed in the external genome rather than in the genome which Locus is a part of. By default None.

_apply_clipped_layer_data(guides, layer_name, layer_data)[source]: Apply layer data to guides.

_get_guide_regions(guide)[source]

_guide_sequence_diversity(guide, g, pos)[source]: Calculate sequence diversity for each region of the guide.

_guide_alt_ac(guide, g, pos)[source]: Calculate alternative allele count for each region of the guide.

_guide_n_variants(guide, g, pos)[source]: Calculate number of variants for each region of the guide.

_apply_variation_layer_data(guides, layer_name, layer_genotype_data, layer_pos)[source]: Apply sequence diversity, alternative allele count and number of variants as layers.

add_layer(name, layer_data, layer_pos=None, apply_to_guides=True, is_variation=False)[source]

Adds a layer with the data to the locus.

Parameters:

namestr: Name of the layer
layer_datanp.ndarray: Layer data. Needs to be the same shape as the locus.
apply_to_guidesbool, optional: Apply layer data to gRNAs when adding it to the locus. By default True.

Examples

>>> locus = Locus("chr1", 100, 200)
>>> layer_data = np.random.rand(100)
>>> locus.add_layer("random", layer_data)

_guide_layers()[source]: Returns a list of all the layers that are present in the gRNAs.

_prepare_alt_matrix(rank_layer_names, method=np.mean)[source]

Prepares numerical matrix with the gRNA layer data to be used later in the ranking.

Parameters:

rank_layer_nameslist: List of layer names to be used in the ranking.
method[type], optional: Method to use to combine the layer data, by default np.mean

Returns:

np.ndarray: Matrix with the layer data for each gRNA.

rank_guides(layer_names=None, layer_is_benefit=None, weight_vector=None, ranking_method='TOPSIS', norm_method='Vector')[source]

Ranks guides based on the layer data.

Returns:

list: List of ranked guides.

guides_to_dataframe()[source]: Returns gRNAs in Pandas dataframe.

guides_to_csv(filename)[source]: Save gRNAs in CSV file.

guides_to_bed(filename)[source]: Save gRNAs in BED file.

guides_detailed_table(filename)[source]: Save gRNAs in a detailed text file.

add_azimuth_score()[source]

Apply Azimuth score to a list of guides.

Azimuth is a machine learning-based predictive modelling of CRISPR/Cas9 guide efficiency. Sometimes its reffered to as Doench 2016 score.

Described in https://doi.org/10.1038/nbt.3437 (Doench et al., 2016)

guido.locus_from_coordinates(genome, chromosome, start, end)[source]

Create a locus from coordinates. Coordinates are 1-based. If annotation file is provided, it will be used to annotate the locus.

Parameters:

genomeGenome: Genome object. Can be created using Genome class.
chromosomestr: Chromosome name.
startint: Start position.
endint: End position.

Returns:

Locus: Locus object.

guido.locus_from_gene(genome, gene_name)[source]

Create a locus from gene name. If annotation file is provided, it will be used to annotate the locus.

Parameters:

genomeGenome: Genome object. Can be created using Genome class.
gene_namestr: Gene name. Needs to be present in the annotation file.

Returns:

Locus: Locus object.

guido.locus_from_sequence(sequence, sequence_name=None)[source]

Create a locus from sequence.

Parameters:

sequencestr: DNA sequence
sequence_namestr, optional: Sequence name, by default None

Returns:

Locus: Object representing a locus from given sequence.

guido

Submodules

Package Contents

Classes

Functions

`guido`