Data types related to simulated populations

Random number generator

class fwdpy11.GSLrng

Random number generator based on GNU Scientific Library mersenne twister.

__init__ = <instancemethod __init__>

Populations of diploids

class fwdpy11.DiploidPopulation(N, length, *, ll_pop=None)

A population of diploid individuals.

To initialize a population of size N and genome length L:

fwdpy11.DiploidPopulation(N, L)

To initialize a population with two demes of initial sizes N1 and N2 and genome length L:

fwdpy11.DiploidPopulation([N1, N2], L)

Changed in version 0.16.0: Added __copy__ and __deepcopy__. In general, the correct way to copy instances if this class is via copy.deepcopy(). Calling copy.copy() will copy some of the underlying C++ objects, but the two objects will share the same table collection. This situation of sharing is almost certainly not what one wants.

generation

mutations

fixations

fixation_times

haploid_genomes

genetic_values

Return the genetic values as a readonly 2d numpy.ndarray.

Rows are individuals. Columns are genetic values.

ancient_sample_genetic_values

Return the genetic values for ancient samples as a readonly 2d numpy.ndarray.

Rows are individuals. Columns are genetic values.

add_mutation(rng, *, window=None, ndescendants=1, deme=None, data=None)

Add a new mutation to the population.

New in version 0.16.0.

Parameters:

rng (fwdpy11.GSLrng) – Random number generator

The following arguments are keyword-only:

Parameters:

• window (tuple[float, float]) – A window [left, right) within which to place the mutation. The default is None, meaning the window is the entire genome.

• ndescendants (int) – The number of alive nodes carrying the new mutation. Default is 1, implying that a singleton mutation will be generated.

• deme (int) – The deme in which to place the new mutation The default is None, meaning that alive node demes are not considered.

Returns:

The key of the new mutation. (The index of the variant in DiploidPopulation.mutations.)

Implementation details:

• A set of nodes are found within window that are ancestral to exactly ndescendants alive nodes.

• From this list of candidate nodes, one is chosen randomly to be the node where we place the new mutation.

• If the node’s parent’s time is NULL, then the mutations’ origin time is the first 64 bit integer ancestral to the node time, which could be the node time itself. Otherwise, a 64 bit integer is chosen uniformly between the node time and the node’s parent’s time.

• If deme is None, any set of ndescendants will be considered.

• If deme is >= 0, all ndescendants alive nodes must be from that deme.

• If the deme/ndescendants requirements cannot be satisified, the function returns None.

• The genetic values of individuals are not updated by this function. Such updates will happen when the population begins evolution forwards in time.

Exceptions:

Raises:

• ValueError – for bad input.

• RuntimeError – if this function is applied during a simulation

• RuntimeError – if internal checks fail.

property alive_nodes

List of alive nodes corresponding to individuals.

New in version 0.5.3.

property ancient_sample_metadata

Supports buffer protocol

property ancient_sample_nodes

Return an array of nodes associated with preserved/ancient samples.

Alias for fwdpy11.DiploidPopulation.preserved_nodes.

New in version 0.13.0.

classmethod create_from_tskit(ts, *, import_mutations=False, import_individuals=False)

Create a new object from an tskit.TreeSequence

Parameters:

• ts (tskit.TreeSequence) – A tree sequence from tskit

• import_mutations (bool) – if True, import mutations and sites from the tree sequence

Returns:

A population object with an initialized fwdpy11.TableCollection

Return type:

fwdpy11.DiploidPopulation

Note

In general, initializing a population using the output from a coalescent simulation is a tricky business. There are issues of parameter scaling and the appropriateness of the coalescent model itself. A key issue is that your input tree sequence must have node times in the correct time units! (Generations, for example.)

Warning

Support for importing mutations is currently limited to two cases:

1. The tree seq has mutations entirely from tskit’s API (see Importing mutations from tskit).

2. The tree seq has mutations entirely from a previous run of fwdpy11.

New in version 0.2.0.

Changed in version 0.19.8: Correctly handle two cases:

1. The tree seq has mutations entirely from tskit’s API (see Importing mutations from tskit).

2. The tree seq has mutations entirely from a previous run of fwdpy11.

deme_sizes(as_dict=False)

Return the number of individuals in each deme.

Parameters:

as_dict – If True, return results as a dict. If False, numpy arrays are returned

The default behavior of this function is equivalent to:

md = numpy.array(self.diploid_metadata, copy=False)
return numpy.unique(md['deme'], return_counts=True)

New in version 0.6.0.

property diploid_metadata

Supports buffer protocol

property diploids

Supports buffer protocol

dump_tables_to_tskit(*, model_params=None, population_metadata=None, data=None, seed=None, parameters=None, destructive=False)

Dump the population’s TableCollection into an tskit TreeSequence

Parameters:

• model_params (fwdpy11.ModelParams) – Model parameters to be stored as top-level metadata

• population_metadata (dict) – A mapping from integer id of a deme/population to metadata

• data (object) – User-generated data to add to top-level metadata

• seed (int) – Random number seed for top-level metadata. If not None, seed must be >= 0.

• parameters (None or dict) – The simulation parameters for the provenance table.

• destructive (bool) – If True, delete data held by the current instance.

Return type:

tskit.TreeSequence

For examples, see Converting data to tskit format.

Warning

If destructive is True, further opertations on the current instance should be considered undefined behavior that may lead to a crash.

Changed in version 0.8.2: Added parameters. Generate provenance information for return value. The provenance information is validated using tskit.validate_provenance(), which may raise an exception.

Changed in version 0.10.0: Use tskit metadata schema. Mutation time is now stored in the tskit.MutationTable column. Origin time of mutations is part of the metadata.

Changed in version 0.14.0: Added destructive option.

Changed in version 0.15.0: Added model_params, demes_graph, population_metadata, data keyword args.

Changed in version 0.15.2: Fixed bug that could generate a tskit.PopulationTable with an incorrect number of rows.

Changed in version 0.17.0: The wrapped keyword argument is deprecated.

New in version 0.19.0: Removed deprecated wrapped keyword argument.

New in version 0.23.0: Remove deprecated demes_graph argument and update type hints.

dump_to_file(filename)

Write a population to a file in binary format.

classmethod load_from_file(filename)

Load population from the output of fwdpy11.DiploidPopulation.dump_to_file().

Parameters:

filename (str) – A file name

classmethod load_from_pickle_file(filename)

Read in a pickled population from a file. The file muse have been generated by a call to fwdpy11.DiploidPopulation.pickle_to_file().

Parameters:

f – A handle to a file opened in ‘rb’ mode.

pickle_to_file(filename)

Pickle the population to an open file.

This function may be preferred over the direct pickling method because it uses less memory. It is, however, slower.

To read the population back in, you must call fwdpy11.DiploidPopulation.load_from_pickle_file().

Parameters:

f – A handle to an open file

New in version 0.3.0.

property preserved_nodes

Return an array of nodes associated with preserved/ancient samples.

Return type:

numpy.ndarray

New in version 0.13.0.

sample_timepoints(include_alive=True)

Return an iterator over all sample time points. The iterator yields time, nodes, and metadata.

Parameters:

include_alive – If True, include currently-alive individuals.

New in version 0.5.1.

Changed in version 0.5.3: Monkey-patched into pybind11 class

property tables

Access the fwdpy11.TableCollection

Diploid Genotypes

This class contains two integers pointing to the genomes making up the individual. The genomes themselves are represented by fwdpy11.HaploidGenome.

class fwdpy11.DiploidGenotype

Class holding indexes of the gametes/genomes making up an individual

first

Key to first gamete. (read-only)

second

Key to second gamete. (read-only)

The diploid genotypes are stored in the following container:

class fwdpy11.DiploidVector

C++ representation of a list of fwdpy11.DiploidGenotype. Typically, access will be read-only.

For all intents and purposes, this container behaves as a standard Python list, but its actual representation is a contiguous array that is handled on the C++ side.

This type supports the Python buffer protocol, meaning that it can be cast to a numpy structured array without a copy, and the attribute names are used as field names.

Diploid metadata

class fwdpy11.DiploidMetadata

Diploid meta data.

property deme

Deme.

property e

Random component of trait value.

property g

Genetic value.

property geography

Array containing the geographic location of the individual.

property label

Index of the individual in the population.

property nodes

Node ids for individual

property parents

Array containing the label fields of the parents.

property sex

Sex.

property w

Fitness.

Metadata are stored in the following container:

class fwdpy11.DiploidMetadataVector

Container of diploid metadata.

For all intents and purposes, this container behaves as a standard Python list, but its actual representation is a contiguous array that is handled on the C++ side.

This type supports the Python buffer protocol, meaning that it can be cast to a numpy structured array without a copy, and the attribute names are used as field names.

HaploidGenomes

Class fwdpy11.HaploidGenome describes a gamete:

class fwdpy11.HaploidGenome

A haploid genome. This object represents the ordered mutations inherited from a single parent.

n

Number of occurrences in the population. This has little meaning beyond book-keeping used by the C++ back-end. (read-only)

mutations

List of keys to neutral mutations. Contains unsigned 32-bit integers corresponding to mutations in the population. (read-only)

smutations

List of keys to selected mutations. Contains unsigned 32-bit integers corresponding to mutations in the population. (read-only)

HaploidGenomes are stored in the following container:

class fwdpy11.HaploidGenomeVector

C++ representations of a list of fwdpy11.HaploidGenome. Typically, access will be read-only.

For all intents and purposes, this container behaves as a standard Python list, but its actual representation is a contiguous array that is handled on the C++ side.

Mutations

A mutation is described by fwdpy11.Mutation:

class fwdpy11.Mutation

Mutation with effect size and dominance

The position of a mutation is a floating-point value:

pos

Position (float).

For simulations with a single effect size/selection coefficient, the value is a float and held in the field:

s

Selection coefficient/effect size. (read-only)

Likewise, the heterozygous effect/dominance of the variant is in the attribute:

h

Dominance/effect in heterozygotes. (read-only)

For simulations with multivariate effects, the analogs of s and h are stored as numpy arrays:

esizes

Vector of effect sizes.

New in version 0.2.0.

Changed in version 0.4.0: Property is now a readonly numpy.ndarray

heffects

Vector of heterozygous effects.

New in version 0.2.0.

Changed in version 0.4.0: Property is now a readonly numpy.ndarray

Other attributes of mutations include:

g

Generation when mutation arose (origination time). (read-only)

label

A 16-bit unsigned integer that can be used for adding “meta-data” to mutations

Note

The label attribute is assigned when a mutation is generated.

Mutations are stored in the following container:

class fwdpy11.MutationVector

C++ representation of a list of fwdpy11.Mutation. Typically, access will be read-only.

For all intents and purposes, this container behaves as a standard Python list, but its actual representation is a contiguous array that is handled on the C++ side.

Types related to tree sequence recording

class fwdpy11.TableCollection(*args)

build_indexes()

Build edge input/output indexes.

New in version 0.13.0.

property edges

Supports buffer protocol

property genome_length

Genome length

property input_left

Edge input order

property mutations

Supports buffer protocol

property nodes

Supports buffer protocol

property output_right

Edge output order

property sites

Supports buffer protocol

class fwdpy11.EdgeTable

An EdgeTable is a container of fwdpy11.Edge.

An EdgeTable supports the Python buffer protocol, allowing data to be viewed in Python as a numpy record array. No copy is made when generating such views.

New in version 0.2.0.

class fwdpy11.Edge

An edge in a tree sequence. An edge represents the transmission of the half-open genomic interval [left,right) from parent to child.

New in version 0.2.0.

parent

Node id of parent

child

Node id of child

left

Left edge of interval, inclusive.

right

Right edge of interval, exclusive.

class fwdpy11.NodeTable

An NodeTable is a container of fwdpy11.Node.

An NodeTable supports the Python buffer protocol, allowing data to be viewed in Python as a numpy record array. No copy is made when generating such views.

New in version 0.2.0.

class fwdpy11.Node

A node in a tree sequence.

New in version 0.2.0.

Changed in version 0.6.0: The population field is renamed deme

deme

For models of discrete population structure, this field is the deme of the node.

time

Birth time of the node, recorded forwards in time.

class fwdpy11.MutationTable

An MutationTable is a container of fwdpy11.MutationRecord.

An MutationTable supports the Python buffer protocol, allowing data to be viewed in Python as a numpy record array. No copy is made when generating such views.

New in version 0.2.0.

class fwdpy11.MutationRecord

A mutation entry in a tree sequence. A MutationRecord stores the node ID of the mutation and the index (“key”) of the mutation in the population.

New in version 0.2.0.

key

Index of the mutation in the population

node

Node id of the mutation

site

Index of the mutation’s site in the population.

New in version 0.5.0.

derived_state

The derived state of the mutation.

New in version 0.5.0.

neutral

Boolean descriptor of whether or not the mutation affects fitness.

New in version 0.5.0.

class fwdpy11.SiteTable

An SiteTable is a container of fwdpy11.Site.

An SiteTable supports the Python buffer protocol, allowing data to be viewed in Python as a numpy record array. No copy is made when generating such views.

New in version 0.5.0.

class fwdpy11.Site

A site in a genome

New in version 0.5.0.

position

The position of the site in the genome

ancestral_state

The ancestral state of the site

class fwdpy11.TreeIterator(tables, samples, *, update_samples=False, ancient_samples=None, begin=0.0, end=None)

Iterate over the marginal trees in a fwdpy11.TableCollection

Positional arguments for initialization:

Parameters:

• tables (TableCollection) – A table collection

• samples (list or numpy.ndarray) – List of sample nodes

The following are keyword only:

Parameters:

• ancient_samples (list or numpy.ndarray) – List of ancient_samples nodes

• begin (float) – Start iteration at this position

• end (float or None) – End iteration at this position

If end is None, then iteration proceeds over all trees with after positions \(\geq \mathrm{begin}\).

If ancient_samples is not None, then TreeIterator.preserved_leaf_counts() will return the number of ancient samples below a node.

Changed in version 0.4.1: Add begin, end options as floats for initializing

leaf_counts(u)

Parameters:

u (int) – A node id

Returns:

number of samples (leaves) below node u

Return type:

int

property left

Left coordinate of current tree (inclusive)

left_child(u)

Parameters:

u (int) – A node id

Returns:

Left child id of node u

Return type:

int

left_sib(u)

Parameters:

u (int) – A node id

Returns:

Left sib id of node u

Return type:

int

mutations()

Return iterator over all fwdpy11.MutationRecord objects on the current tree.

New in version 0.5.1.

nodes()

Return the nodes in the current tree.

The return order is preorder.

New in version 0.5.1.

parent(u)

Parameters:

u (int) – A node id

Returns:

Parent id of node u

Return type:

int

preserved_leaf_counts(u)

Parameters:

u (int) – A node id

Returns:

number of ancient samples below node u

Return type:

int

property right

Right coordinate of current tree (exclusive).

right_child(u)

Parameters:

u (int) – A node id

Returns:

Right child id of node u

Return type:

int

right_sib(u)

Parameters:

u (int) – A node id

Returns:

Right sib id of node u

Return type:

int

property roots

Roots of the current tree

samples()

Returns:

The samples list

Return type:

numpy.ndarray

samples_below(node, sort=False)

Return the list of samples descending from a node.

Parameters:

• node (int) – A node id

• sort (bool) – (False) Whether or not to sort sample node IDs.

Note

Do not store these sample lists without making a “deep” copy. The internal buffer is re-used.

sites()

Return iterator over all fwdpy11.Site objects on the current tree.

New in version 0.5.1.

property tables

The underlying TableCollection

total_time()

Returns:

Sum of branch lengths

Return type:

float

class fwdpy11.VariantIterator(tables, samples, *, begin=0.0, end=None, include_neutral_variants=True, include_selected_variants=True)

An iterable class for traversing genotypes in a tree sequence.

To initialize:

Parameters:

• tables (fwdpy11.TableCollection) – The table collection

• samples (list or numpy.ndarray) – Samples list

The following are kwarg-only:

Parameters:

• begin (float) – (0.0) First position, inclusive.

• end (float) – (None) Last position, exclusive.

• include_neutral_variants (bool) – (True) Include neutral variants during traversal

• include_selected_variants (bool) – (True) Include selected variants during traversal

Changed in version 0.4.1: Add begin, end options as floats

Changed in version 0.4.2: Add include_neutral_variants and include_selected_variants

Changed in version 0.5.0: No longer requires a fwdpy11.MutationVector.

Changed in version 0.14.0: Initialization now requires a TableCollection. Some initialization parmeters are now keyword-only.

property genotypes

Genotype array. Index order is same as sample input

property position

Current mutation position

property records

fwdpy11.MutationRecord objects for the current Site.

property site

The current fwdpy11.Site

class fwdpy11.DataMatrixIterator(tables, samples, intervals, neutral, selected, fixations=False)

Efficient iteration over genomic windows.

This class allows efficient traversal across multiple genomic intervals. The class is iterable, and encapsulates the data fields of fwdpy11.DataMatrix, which are accessible as properties.

To initialize:

Parameters:

• tables (fwdpy11.TableCollection) – A table collection

• samples (list or numpy.ndarray) – A list of samples

• intervals (list[tuple]) – The \([start, stop)\) positions of each interval

• neutral (bool) – If True, include neutral variants

• selected (bool) – If True, include selected variants

• fixations (bool) – (False) If True, include fixations in the sample

New in version 0.4.4.

Changed in version 0.5.0: Initialization longer requires fwdpy11.MutationVector

The class constructor is:

__init__ = <function DataMatrixIterator.__init__>

The following properties are numpy arrays providing read-only access to an internal instance of fwdpy11.DataMatrix:

neutral

Genotypes at neutral variants

selected

Genotypes at selected variants.

neutral_keys

Indexes of the neutral variants.

selected_keys

Indexes of selected variants.

neutral_positions

Positions of neutral variants.

selected_positions

Positions of selected variants.

class fwdpy11.NoAncientSamples

A recorder for tree sequence simulations that does nothing.

__init__ = <instancemethod __init__>

Types related to discrete demographic events

class fwdpy11.ForwardTimeInterval(start_time, end_time)

A half-open interval [start_time, end_time). The times represent values moving forward in time, in generations since the begining of a model.

class fwdpy11.ForwardDemesGraph(yaml, burnin, burnin_is_exact, round_non_integer_sizes)

A forward-in-time representation of a demes.Graph

Parameters:

• yaml (str) – The string representation of a demes model.

• burnin (int) – The number of generations of evolution to occur before the events in the yaml take place.

• burnin_is_exact (bool) – If False, burnin will be treated as an integer multiple of the sum of ancestor population sizes. If True, burnin will be used as-is.

• round_non_integer_sizes (bool) – If True, a graph containing non-integer values for epoch start and/or end sizes will have those values rounded to the nearest integer.

The recommended method for construction is ForwardDemesGraph.from_demes() or ForwardDemesGraph.tubes().

New in version 0.20.0.

asdict()

Return dict representation

property burnin_duration

The length of the burning time. Equal to burnin_generation + 1.

property deme_labels

A dictionary mapping integer deme IDs to deme names (strings).

deme_time_intervals()

Returns:

a list of the time intervals during which each deme exists.

Return type:

list[fwdpy11.ForwardTimeInterval]

The order of the demes is the same as in the input demes.Graph.

property demes_at_final_generation

List of integer ids of extant demes in the final generation of the demes graph.

property demes_graph

Obtain the internal representation of the demes.Graph

This object is not necessarily equivalent to the input graph:

• It is fully resolved.

• It has been rounded to discrete time and deme sizes.

epoch_time_intervals()

Returns:

a list of the time intervals during which each epoch of each deme exists.

Return type:

list[list[fwdpy11.ForwardTimeInterval]]

The order of the demes is the same as in the input demes.Graph. Within each deme, the epoch order is the same as in the input graph.

property final_generation

This is the final offspring generation of the model. In other words, a model evolved from parental generation zero until time zero of the graph (backwards in time) will result in individuals whose birth times equal the value of this property.

classmethod from_demes(model, burnin, *, burnin_is_exact=None, round_non_integer_sizes=None)

Build from a demes graph.

Parameters:

• model (str or demes.Graph) – A demes model.

• burnin (int) – The number of generations of evolution to occur before the events in the yaml take place.

• burnin_is_exact (bool) – If False, burnin will be treated as an integer multiple of the sum of ancestor population sizes. If True, burnin will be used as-is.

Returns:

A forward-time representation of a demes graph

Return type:

fwdpy11.ForwardDemesGraph

classmethod fromdict(d)

Build an instance from a dictionary

property initial_sizes

A list of the nonzero (parental) deme sizes at time 0 of the model (thinking forwards in time).

Due to the structure of a demes graph, the indexes of the list are the deme integer identifiers.

This property is useful for setting up a fwdpy11.DiploidPopulation.

property model_duration

The duration of the model from time 0 until the most recent time point.

to_backwards_time(forwards_time)

Convert an integer value representing time moving in the forward direction to time in the past (“backwards time”).

If no conversion is possible, None is returned.

Parameters:

forwards_time (int) – The value to convert

Returns:

The converted time

Type:

Optional[int]

to_forwards_time(backwards_time)

Convert an integer value representing time moving in the backward direction to time in the forward direction.

If no conversion is possible, None is returned.

Parameters:

backwards_time (int) – The value to convert

Returns:

The converted time

Type:

Optional[int]

classmethod tubes(sizes, burnin, *, burnin_is_exact=None, round_non_integer_sizes=None)

Build a demographic model for one or more demes of constant size.

Parameters:

• sizes (List[int]) – Deme sizes

• burnin (int) – The number of generations of evolution to simulate.

• burnin_is_exact (bool) – If False, burnin will be treated as an integer multiple of the sum of ancestor population sizes. If True, burnin will be used as-is.

Returns:

A forward-time representation of a demes graph

Return type:

fwdpy11.ForwardDemesGraph

class fwdpy11.DemographyDebugger(initial_deme_sizes, events, simlen=None, deme_labels=None)

Efficiently debug demographic events.

The class executes a mock simulation of demographic events in order to quickly detect errors. It also generates a “report” of the model details for double-checking.

If model errors are found, a ValueError exception is raised. The goal is to catch the types of errors that would result in an exception during a simulation.

The class is initialized with the following arguments, which may be either positional or keyword:

Parameters:

• initial_deme_sizes (list or fwdpy11.DiploidPopulation) – The initial sizes of each deme

• events (fwdpy11.ForwardDemesGraph or fwdpy11.DemographicModelDetails) – The demographic model

• simlen (int) – The length of the simulation. Defaults to None.

• deme_labels (dict) – A map from deme index to a printable name

New in version 0.6.0.

Changed in version 0.8.1: Initialization now done with attr. A list of initial deme sizes is now accepted.

class fwdpy11.DemographicModelDetails(*args, **kwargs)

Stores rich information about a demographic model. Instances of this class get returned by functions generating pre-calculated models.

This class has the following attributes, whose names are also kwargs for intitialization. The attribute names also determine the order of positional arguments:

Parameters:

• model (fwdpy11.ForwardDemesGraph) – The demographic model parameters

• name (str) – The name of the model

• source (dict) – The source of the model

• parameters (object) – The parameters used to generate model

• citation (dict or fwdpy11.DemographicModelCitation or None) – Citation information

• metadata (object) – Optional field for additional info

New in version 0.8.0.

asblack()

Return a string representation formatted with black

asdict()

Return dict representation

property deme_labels

A dictionary mapping integer deme IDs to deme names (strings).

classmethod fromdict(d)

Build an instance from a dictionary

property initial_sizes_list

A list of the nonzero (parental) deme sizes at time 0 of the model (thinking forwards in time).

This property is useful for setting up a fwdpy11.DiploidPopulation.

property total_simulation_length

The total number of generations in the model.

This value includes any burn-in time and starts from time 0 (forwards in time).

class fwdpy11.DemographicModelCitation(*args, **kwargs)

Citation information for a demographic model

This class has the following attributes, whose names are also kwargs for intitialization. The attribute names also determine the order of positional arguments:

Parameters:

• DOI – The Digital Object Identifier

• full_citation – Something string-like giving the full citation.

• metadata – Any additional information that may be needed.

New in version 0.8.0.

Miscellaneous types

class fwdpy11.RecordNothing

__init__ = <instancemethod __init__>

class fwdpy11.NewMutationData(*, effect_size, dominance, esizes=None, heffects=None, label=0)

Data object for fwdpy11.DiploidPopulation.add_mutation()

New in version 0.16.0.

This class has the following attributes, whose names are also kwargs for intitialization. The attribute names also determine the order of positional arguments:

Parameters:

• effect_size (float) – The fixed/univariate effect size. Will fill in fwdpy11.Mutation.s

• dominance (float) – Heterozygous effect of the mutation. Will fill in fwdpy11.Mutation.h

• esizes (list[float]) – Data to fill fwdpy11.Mutation.esizes. Default is None.

• heffects (list[float]) – Data to fill fwdpy11.Mutation.heffects. Default is None.

• label (int) – Data for fwdpy11.Mutation.label. Default is 0.