biotite 0.41.1__cp311-cp311-macosx_10_16_x86_64.whl

biotite/sequence/annotation.py

@@ -0,0 +1,829 @@
+# This source code is part of the Biotite package and is distributed
+# under the 3-Clause BSD License. Please see 'LICENSE.rst' for further
+# information.
+
+__name__ = "biotite.sequence"
+__author__ = "Patrick Kunzmann"
+__all__ = ["Location", "Feature", "Annotation", "AnnotatedSequence"]
+
+import numbers
+import copy
+import sys
+from enum import Flag, Enum, auto
+import numpy as np
+from .sequence import Sequence
+from ..copyable import Copyable
+from .seqtypes import NucleotideSequence
+
+
+class Location():
+    """
+    A :class:`Location` defines at which base(s)/residue(s) a feature is
+    located.
+
+    A feature can have multiple :class:`Location` instances if multiple
+    locations are joined.
+
+    Objects of this class are immutable.
+
+    Attributes
+    ----------
+    first : int
+        Starting base or residue position of the feature.
+    last : int
+        Inclusive ending base or residue position of the feature.
+    strand : Strand
+        The strand direction.
+        Always :attr:`Strand.FORWARD` for peptide features.
+    defect : Defect
+        A possible defect of the location.
+    """
+
+    class Defect(Flag):
+        """
+        This enum type describes location defects.
+
+        A location has a defect, when the feature itself is not directly
+        located in the range of the first to the last base.
+
+           - **NONE** - No location defect
+           - **MISS_LEFT** - A part of the feature has been truncated
+             before the first base/residue of the :class:`Location`
+             (probably by indexing an :class:`Annotation` object)
+           - **MISS_RIGHT** - A part of the feature has been truncated
+             after the last base/residue of the :class:`Location`
+             (probably by indexing an :class:`Annotation` object)
+           - **BEYOND_LEFT** - The feature starts at an unknown position
+             before the first base/residue of the :class:`Location`
+           - **BEYOND_RIGHT** - The feature ends at an unknown position
+             after the last base/residue of the :class:`Location`
+           - **UNK_LOC** - The exact position is unknown, but it is at a
+             single base/residue between the first and last residue of
+             the :class:`Location`, inclusive
+           - **BETWEEN** - The position is between to consecutive
+             bases/residues.
+        """
+        NONE         = 0
+        MISS_LEFT    = auto()
+        MISS_RIGHT   = auto()
+        BEYOND_LEFT  = auto()
+        BEYOND_RIGHT = auto()
+        UNK_LOC      = auto()
+        BETWEEN      = auto()
+
+    class Strand(Enum):
+        """
+        This enum type describes the strand of the feature location.
+        This is not relevant for protein sequence features.
+        """
+        FORWARD = auto()
+        REVERSE = auto()
+
+    def __init__(self, first, last, strand=Strand.FORWARD,
+                 defect=Defect.NONE):
+        if first > last:
+            raise ValueError(
+                "The first position cannot be higher than the last position"
+            )
+        self._first = first
+        self._last = last
+        self._strand = strand
+        self._defect = defect
+
+    def __repr__(self):
+        """Represent Location as a string for debugging."""
+        return f'Location({self._first}, {self._last}, strand={"Location." + str(self._strand)}, ' \
+               f'defect={"Location." + str(self._defect)})'
+
+    @property
+    def first(self):
+        return self._first
+
+    @property
+    def last(self):
+        return self._last
+
+    @property
+    def strand(self):
+        return self._strand
+
+    @property
+    def defect(self):
+        return self._defect
+
+    def __str__(self):
+        string = "{:d}-{:d}".format(self.first, self.last)
+        if self.strand == Location.Strand.FORWARD:
+            string = string + " >"
+        else:
+            string = "< " + string
+        return string
+
+    def __eq__(self, item):
+        if not isinstance(item, Location):
+            return False
+        return (    self.first  == item.first
+                and self.last   == item.last
+                and self.strand == item.strand
+                and self.defect == item.defect)
+
+    def __hash__(self):
+        return hash((self._first, self._last, self._strand, self._defect))
+
+
+class Feature(Copyable):
+    """
+    This class represents a single sequence feature, for example from a
+    GenBank feature table.
+    A feature describes a functional part of a sequence.
+    It consists of a feature key, describing the general class of the
+    feature, at least one location, describing its position on the
+    reference, and qualifiers, describing the feature in detail.
+
+    Objects of this class are immutable.
+
+    Parameters
+    ----------
+    key : str
+        The name of the feature class, e.g. *gene*, *CDS* or
+        *regulatory*.
+    locs : iterable object of Location
+        A list of feature locations. In most cases this list will only
+        contain one location, but multiple ones are also possible for
+        example in eukaryotic CDS (due to splicing).
+    qual : dict, optional
+        Maps feature qualifiers to their corresponding values.
+        The keys are always strings. A value is either a string or
+        ``None`` if the qualifier key do not has a value.
+        If key has multiple values, the values are separated by a
+        line break.
+
+    Attributes
+    ----------
+    key : str
+        The name of the feature class, e.g. *gene*, *CDS* or
+        *regulatory*.
+    locs : iterable object of Location
+        A list of feature locations. In most cases this list will only
+        contain one location, but multiple ones are also possible for
+        example in eukaryotic CDS (due to splicing).
+    qual : dict
+        Maps feature qualifiers to their corresponding values.
+        The keys are always strings. A value is either a string or
+        ``None`` if the qualifier key do not has a value.
+        If key has multiple values, the values are separated by a
+        line break.
+    """
+
+    def __init__(self, key, locs, qual=None):
+        self._key = key
+        if len(locs) == 0:
+            raise ValueError("A feature must have at least one location")
+        self._locs = frozenset(locs)
+        self._qual = copy.deepcopy(qual) if qual is not None else {}
+
+    def __repr__(self):
+        """Represent Feature as a string for debugging."""
+        return f'Feature("{self._key}", [{", ".join([loc.__repr__() for loc in self.locs])}], qual={self._qual})'
+
+    def get_location_range(self):
+        """
+        Get the minimum first base/residue and maximum last base/residue
+        of all feature locations.
+
+        This can be used to create a location, that spans all of the
+        feature's locations.
+
+        Returns
+        -------
+        first : int
+            The minimum first base/residue of all locations.
+        last : int
+            The maximum last base/residue of all locations.
+        """
+        first = np.min([loc.first for loc in self._locs])
+        last = np.max([loc.last for loc in self._locs])
+        return first, last
+
+    def __eq__(self, item):
+        if not isinstance(item, Feature):
+            return False
+        return (    self._key  == item._key
+                and self._locs == item._locs
+                and self._qual == item._qual)
+
+    def __lt__(self, item):
+        if not isinstance(item, Feature):
+            return False
+        first, last = self.get_location_range()
+        it_first, it_last = item.get_location_range()
+        # The first base/residue is most significant,
+        # if it is equal for both features, look at last base/residue
+        if first < it_first:
+            return True
+        elif first > it_first:
+            return False
+        else: # First is equal
+            return last > it_last
+
+    def __gt__(self, item):
+        if not isinstance(item, Feature):
+            return False
+        first, last = self.get_location_range()
+        it_first, it_last = item.get_location_range()
+        # The first base/residue is most significant,
+        # if it is equal for both features, look at last base/residue
+        if first > it_first:
+            return True
+        elif first < it_first:
+            return False
+        else: # First is equal
+            return last < it_last
+
+    @property
+    def key(self):
+        return self._key
+
+    @property
+    def locs(self):
+        return copy.copy(self._locs)
+
+    @property
+    def qual(self):
+        return copy.copy(self._qual)
+
+    def __hash__(self):
+        return hash(( self._key, self._locs, frozenset(self._qual.items()) ))
+
+
+class Annotation(Copyable):
+    """
+    An :class:`Annotation` is a set of features belonging to one
+    sequence.
+
+    Its advantage over a simple list is the base/residue position based
+    indexing:
+    When using slice indices in Annotation objects, a subannotation is
+    created, containing copies of all :class:`Feature` objects whose
+    first and last base/residue are in range of the slice.
+    If the slice starts after the first base/residue or/and the slice
+    ends before the last residue, the position out of range is set to
+    the boundaries of the slice (the :class:`Feature` is truncated).
+    In this case the :class:`Feature` obtains the
+    :attr:`Location.Defect.MISS_LEFT` and/or
+    :attr:`Location.Defect.MISS_RIGHT` defect.
+    The third case occurs when a :class:`Feature` starts after the slice
+    ends or a :class:`Feature` ends before the slice starts.
+    In this case the :class:`Feature` will not appear in the
+    subannotation.
+
+    The start or stop position in the slice indices can be omitted, then
+    the subannotation will include all features from the start or up to
+    the stop, respectively. Step values are ignored.
+    The stop values are still exclusive, i.e. the subannotation will
+    contain a not truncated :class:`Feature` only if its last
+    base/residue is smaller than the stop value of the slice.
+
+    Integers or other index types are not supported. If you want to
+    obtain the :class:`Feature` instances from the :class:`Annotation`
+    you need to  iterate over it.
+    The iteration has no defined order.
+    Alternatively, you can obtain a copy of the internal
+    :class:`Feature` set via :func:`get_features()`.
+
+    Multiple :class:`Annotation` objects can be concatenated to one
+    :class:`Annotation` object using the '+' operator.
+    Single :class:`Feature` instances can be added this way, too.
+    If a feature is present in both :class:`Annotation` objects, the
+    resulting :class:`Annotation` will contain this feature twice.
+
+    Parameters
+    ----------
+    features : iterable object of Feature, optional
+        The features to create the :class:`Annotation` from. if not
+        provided, an empty :class:`Annotation` is created.
+
+    Examples
+    --------
+    Creating an annotation from a feature list:
+
+    >>> feature1 = Feature("CDS", [Location(-10, 30 )], qual={"gene" : "test1"})
+    >>> feature2 = Feature("CDS", [Location(20,  50 )], qual={"gene" : "test2"})
+    >>> annotation = Annotation([feature1, feature2])
+    >>> for f in sorted(list(annotation)):
+    ...     print(f.qual["gene"], "".join([str(loc) for loc in f.locs]))
+    test1 -10-30 >
+    test2 20-50 >
+
+    Merging two annotations and a feature:
+
+    >>> feature3 = Feature("CDS", [Location(100, 130 )], qual={"gene" : "test3"})
+    >>> feature4 = Feature("CDS", [Location(150, 250 )], qual={"gene" : "test4"})
+    >>> annotation2 = Annotation([feature3, feature4])
+    >>> feature5 = Feature("CDS", [Location(-50, 200 )], qual={"gene" : "test5"})
+    >>> annotation = annotation + annotation2 + feature5
+    >>> for f in sorted(list(annotation)):
+    ...     print(f.qual["gene"], "".join([str(loc) for loc in f.locs]))
+    test5 -50-200 >
+    test1 -10-30 >
+    test2 20-50 >
+    test3 100-130 >
+    test4 150-250 >
+
+    Location based indexing, note the defects:
+
+    >>> annotation = annotation[40:150]
+    >>> for f in sorted(list(annotation)):
+    ...     gene = f.qual["gene"]
+    ...     loc_str = "".join([f"{loc}    {loc.defect}" for loc in f.locs])
+    ...     print(gene, loc_str)
+    test5 40-149 >    Defect.MISS_RIGHT|MISS_LEFT
+    test2 40-50 >    Defect.MISS_LEFT
+    test3 100-130 >    Defect.NONE
+    """
+
+    def __init__(self, features=None):
+        if features is None:
+            self._features = set()
+        else:
+            self._features = set(features)
+
+    def __repr__(self):
+        """Represent Annotation as a string for debugging."""
+        return f'Annotation([{", ".join([feat.__repr__() for feat in self._features])}])'
+
+    def __copy_create__(self):
+        return Annotation(self._features)
+
+    def get_features(self):
+        """
+        Get a copy of the internal feature set.
+
+        Returns
+        -------
+        feature_list : list of Feature
+            A copy of the internal feature set.
+        """
+        return copy.copy(self._features)
+
+    def add_feature(self, feature):
+        """
+        Add a feature to the annotation.
+
+        Parameters
+        ----------
+        feature : Feature
+            Feature to be added.
+        """
+        if not isinstance(feature, Feature):
+            raise TypeError(
+                f"Only 'Feature' objects are supported, "
+                f"not {type(feature).__name__}"
+            )
+        self._features.add(feature)
+
+    def get_location_range(self):
+        """
+        Get the range of feature locations,
+        i.e. the first and exclusive last base/residue.
+
+        Returns
+        -------
+        int : start
+            Start location.
+        int : stop
+            Exclusive stop location.
+        """
+        first = sys.maxsize
+        last = -sys.maxsize
+        for feature in self._features:
+            for loc in feature.locs:
+                if loc.first < first:
+                    first = loc.first
+                if loc.last > last:
+                    last = loc.last
+        # Exclusive stop -> +1
+        return first, last+1
+
+    def del_feature(self, feature):
+        """
+        Delete a feature from the annotation.
+
+        Parameters
+        ----------
+        feature : Feature
+            Feature to be removed.
+
+        Raises
+        ------
+        KeyError
+            If the feature is not in the annotation
+        """
+        self._features.remove(feature)
+
+    def __add__(self, item):
+        if isinstance(item, Annotation):
+            return Annotation(self._features | item._features)
+        elif isinstance(item, Feature):
+            return Annotation(self._features | set([item]))
+        else:
+            raise TypeError(
+                f"Only 'Feature' and 'Annotation' objects are supported, "
+                f"not {type(item).__name__}"
+            )
+
+    def __iadd__(self, item):
+        if isinstance(item, Annotation):
+            self._features |= item._features
+        elif isinstance(item, Feature):
+            self._features.add(item)
+        else:
+            raise TypeError(
+                f"Only 'Feature' and 'Annotation' objects are supported, "
+                f"not {type(item).__name__}"
+            )
+        return self
+
+    def __getitem__(self, index):
+        if isinstance(index, slice):
+            # If no start or stop index is given, include all
+            if index.start is None:
+                i_first = -sys.maxsize
+            else:
+                i_first = index.start
+            if index.stop is None:
+                i_last = sys.maxsize
+            else:
+                i_last = index.stop - 1
+
+            sub_annot = Annotation()
+            for feature in self:
+                locs_in_scope = []
+                for loc in feature.locs:
+                    # Always true for maxsize values
+                    # in case no start or stop index is given
+                    if loc.first <= i_last and loc.last >= i_first:
+                        # The location is at least partly in the
+                        # given location range
+                        # Handle defects
+                        first = loc.first
+                        last = loc.last
+                        defect = loc.defect
+                        if loc.first < i_first:
+                            defect |= Location.Defect.MISS_LEFT
+                            first = i_first
+                        if loc.last > i_last:
+                            defect |= Location.Defect.MISS_RIGHT
+                            last = i_last
+                        locs_in_scope.append(Location(
+                            first, last, loc.strand, defect
+                        ))
+                if len(locs_in_scope) > 0:
+                    # The feature is present in the new annotation
+                    # if any of the original locations is in the new
+                    # scope
+                    new_feature = Feature(
+                        key=feature.key, locs=locs_in_scope, qual=feature.qual
+                    )
+                    sub_annot.add_feature(new_feature)
+            return sub_annot
+        else:
+            raise TypeError(
+                f"'{type(index).__name__}' instances are invalid indices"
+            )
+
+    def __delitem__(self, item):
+        if not isinstance(item, Feature):
+            raise TypeError(
+                f"Only 'Feature' objects are supported, "
+                f"not {type(item).__name__}"
+            )
+        self.del_feature(item)
+
+    def __iter__(self):
+        return self._features.__iter__()
+
+    def __contains__(self, item):
+        return item in self._features
+
+    def __eq__(self, item):
+        if not isinstance(item, Annotation):
+            return False
+        return self._features == item._features
+
+    def __len__(self):
+        return len(self._features)
+
+
+class AnnotatedSequence(Copyable):
+    """
+    An :class:`AnnotatedSequence` is a combination of a
+    :class:`Sequence` and an :class:`Annotation`.
+
+    Indexing an :class:`AnnotatedSequence` with a slice returns another
+    :class:`AnnotatedSequence` with the corresponding subannotation and
+    a sequence start corrected subsequence, i.e. indexing starts at 1
+    with the default sequence start 1.
+    The sequence start in the newly created :class:`AnnotatedSequence`
+    is the start of the slice.
+    Furthermore, integer indices are allowed in which case the
+    corresponding symbol of the sequence is returned (also sequence
+    start corrected).
+    In both cases the index must be in range of the sequence, e.g. if
+    sequence start is 1, index 0 is not allowed.
+    Negative indices do not mean indexing from the end of the sequence,
+    in contrast to the behavior in :class:`Sequence` objects.
+    Both index types can also be used to modify the sequence.
+
+    Another option is indexing with a :class:`Feature` (preferably from the
+    :class:`Annotation` in the same :class:`AnnotatedSequence`).
+    In this case a sequence, described by the location(s) of the
+    :class:`Feature`, is returned.
+    When using a :class:`Feature` for setting an
+    :class:`AnnotatedSequence` with a sequence, the new sequence is
+    replacing the locations of the
+    :class:`Feature`.
+    Note the the replacing sequence must have the same length as the
+    sequence of the :class:`Feature` index.
+
+    Parameters
+    ----------
+    sequence : Sequence
+        The sequence.
+        Usually a :class:`NucleotideSequence` or
+        :class:`ProteinSequence`.
+    annotation : Annotation
+        The annotation corresponding to `sequence`.
+    sequence_start : int, optional
+        By default, the first symbol of the sequence is corresponding
+        to location 1 of the features in the annotation. The location
+        of the first symbol can be changed by setting this parameter.
+        Negative values are not supported yet.
+
+    Attributes
+    ----------
+    sequence : Sequence
+        The represented sequence.
+    annotation : Annotation
+        The annotation corresponding to `sequence`.
+    sequence_start : int
+        The location of the first symbol in the sequence.
+
+    See also
+    --------
+    Annotation, Sequence
+
+    Examples
+    --------
+    Creating an annotated sequence
+
+    >>> sequence = NucleotideSequence("ATGGCGTACGATTAGAAAAAAA")
+    >>> feature1 = Feature("misc_feature", [Location(1,2), Location(11,12)],
+    ...                    {"note" : "walker"})
+    >>> feature2 = Feature("misc_feature", [Location(16,22)], {"note" : "poly-A"})
+    >>> annotation = Annotation([feature1, feature2])
+    >>> annot_seq = AnnotatedSequence(annotation, sequence)
+    >>> print(annot_seq.sequence)
+    ATGGCGTACGATTAGAAAAAAA
+    >>> for f in sorted(list(annot_seq.annotation)):
+    ...     print(f.qual["note"])
+    walker
+    poly-A
+
+    Indexing with integers, note the sequence start correction
+
+    >>> print(annot_seq[2])
+    T
+    >>> print(annot_seq.sequence[2])
+    G
+
+    indexing with slices
+
+    >>> annot_seq2 = annot_seq[:16]
+    >>> print(annot_seq2.sequence)
+    ATGGCGTACGATTAG
+    >>> for f in annot_seq2.annotation:
+    ...     print(f.qual["note"])
+    walker
+
+    Indexing with features
+
+    >>> print(annot_seq[feature1])
+    ATAT
+    >>> print(annot_seq[feature2])
+    AAAAAAA
+    >>> print(annot_seq.sequence)
+    ATGGCGTACGATTAGAAAAAAA
+    >>> annot_seq[feature1] = NucleotideSequence("CCCC")
+    >>> print(annot_seq.sequence)
+    CCGGCGTACGCCTAGAAAAAAA
+    """
+
+    def __init__(self, annotation, sequence, sequence_start=1):
+        self._annotation = annotation
+        self._sequence = sequence
+        self._seqstart = sequence_start
+
+    def __repr__(self):
+        """Represent AnnotatedSequence as a string for debugging."""
+        return f'AnnotatedSequence({self._annotation.__repr__()}, {self._sequence.__repr__()}, ' \
+               f'sequence_start={self._seqstart})'
+
+    @property
+    def sequence_start(self):
+        return self._seqstart
+
+    @property
+    def sequence(self):
+        return self._sequence
+
+    @property
+    def annotation(self):
+        return self._annotation
+
+    def __copy_create__(self):
+        return AnnotatedSequence(
+            self._annotation.copy(), self._sequence.copy, self._seqstart)
+
+    def reverse_complement(self, sequence_start=1):
+        """
+        Create the reverse complement of the annotated sequence.
+
+        This method accurately converts the position and the strand of
+        the annotation.
+        The information on the sequence start is lost.
+
+        Parameters
+        ----------
+        sequence_start : int, optional
+            The location of the first symbol in the reverse complement
+            sequence.
+
+        Returns
+        -------
+            The reverse complement of the annotated sequence.
+        """
+        rev_seqstart = sequence_start
+
+        rev_sequence = self._sequence.reverse().complement()
+
+        seq_len = len(self._sequence)
+        rev_features = []
+        for feature in self._annotation:
+            rev_locs = []
+            for loc in feature.locs:
+                # Transform location to the reverse complement strand
+                # (seq_len-1) -> last sequence index
+                # (loc.last-self._seqstart) -> location to index
+                # ... + rev_seqstart -> index to location
+                rev_loc_first \
+                    = (seq_len-1) - (loc.last-self._seqstart) + rev_seqstart
+                rev_loc_last \
+                    = (seq_len-1) - (loc.first-self._seqstart) + rev_seqstart
+
+                if loc.strand == Location.Strand.FORWARD:
+                    rev_loc_strand = Location.Strand.REVERSE
+                else:
+                    rev_loc_strand = Location.Strand.FORWARD
+
+                rev_loc_defect = Location.Defect.NONE
+                if loc.defect & Location.Defect.MISS_LEFT:
+                    rev_loc_defect |= Location.Defect.MISS_RIGHT
+                if loc.defect & Location.Defect.MISS_RIGHT:
+                    rev_loc_defect |= Location.Defect.MISS_LEFT
+                if loc.defect & Location.Defect.BEYOND_RIGHT:
+                    rev_loc_defect |= Location.Defect.BEYOND_LEFT
+                if loc.defect & Location.Defect.BEYOND_LEFT:
+                    rev_loc_defect |= Location.Defect.BEYOND_RIGHT
+                if loc.defect & Location.Defect.UNK_LOC:
+                    rev_loc_defect |= Location.Defect.UNK_LOC
+                if loc.defect & Location.Defect.BETWEEN:
+                    rev_loc_defect |= Location.Defect.BETWEEN
+
+                rev_locs.append(Location(
+                        rev_loc_first, rev_loc_last,
+                        rev_loc_strand, rev_loc_defect
+                ))
+            rev_features.append(Feature(
+                feature.key, rev_locs, feature.qual
+            ))
+
+        return AnnotatedSequence(
+            Annotation(rev_features), rev_sequence, rev_seqstart
+        )
+
+    def __getitem__(self, index):
+        if isinstance(index, Feature):
+            # Concatenate subsequences for each location of the feature
+            locs = index.locs
+            if len(locs) == 0:
+                raise ValueError("Feature does not contain any locations")
+            # Start by creating an empty sequence
+            sub_seq = self._sequence.copy(new_seq_code=np.array([]))
+            # Locations need to be sorted, as otherwise the locations
+            # chunks would be merged in the wrong order
+            # The order depends on whether the locs are on the forward
+            # or reverse strand
+            strand = None
+            for loc in locs:
+                if loc.strand == strand:
+                    pass
+                elif strand is None:
+                    strand = loc.strand
+                else: # loc.strand != strand
+                    raise ValueError(
+                        "All locations of the feature must have the same "
+                        "strand direction"
+                    )
+            if strand == Location.Strand.FORWARD:
+                sorted_locs = sorted(
+                    locs, key=lambda loc: loc.first
+                )
+            else:
+                sorted_locs = sorted(
+                    locs, key=lambda loc: loc.last, reverse=True
+                )
+            # Merge the sequences corresponding to the ordered locations
+            for loc in sorted_locs:
+                slice_start = loc.first - self._seqstart
+                # +1 due to exclusive stop
+                slice_stop = loc.last - self._seqstart +1
+                add_seq = self._sequence[slice_start:slice_stop]
+                if loc.strand == Location.Strand.REVERSE:
+                    add_seq = add_seq.reverse().complement()
+                sub_seq += add_seq
+            return sub_seq
+
+        elif isinstance(index, slice):
+            # Sequence start correction
+            if index.start is None:
+                seq_start = 0
+            else:
+                if index.start < self._seqstart:
+                    raise IndexError(
+                        f"The start of the index ({index.start}) is lower "
+                        f"than the start of the sequence ({self._seqstart})"
+                    )
+                seq_start = index.start - self._seqstart
+            if index.stop is None:
+                seq_stop = len(self._sequence)
+                index = slice(index.start, seq_stop, index.step)
+            else:
+                seq_stop = index.stop - self._seqstart
+            # New value for the sequence start, value is base position
+            if index.start is None:
+                rel_seq_start = self._seqstart
+            else:
+                rel_seq_start = index.start
+            return AnnotatedSequence(self._annotation[index],
+                                     self._sequence[seq_start:seq_stop],
+                                     rel_seq_start)
+
+        elif isinstance(index, numbers.Integral):
+            return self._sequence[index - self._seqstart]
+
+        else:
+            raise TypeError(
+                f"'{type(index).__name__}' instances are invalid indices"
+            )
+
+    def __setitem__(self, index, item):
+        if isinstance(index, Feature):
+            # Item must be sequence
+            # with length equal to sum of location lengths
+            sub_seq = item
+            sub_seq_i = 0
+            for loc in index.locs:
+                slice_start = loc.first - self._seqstart
+                # +1 due to exclusive stop
+                slice_stop = loc.last - self._seqstart +1
+                interval_size = slice_stop - slice_start
+                self._sequence[slice_start:slice_stop] \
+                    = sub_seq[sub_seq_i : sub_seq_i + interval_size]
+                sub_seq_i += interval_size
+        elif isinstance(index, slice):
+            # Sequence start correction
+            if index.start is None:
+                seq_start = 0
+            else:
+                seq_start = index.start - self._seqstart
+            if index.stop is None:
+                seq_stop = len(self._sequence)
+            else:
+                seq_stop = index.stop - self._seqstart
+            # Item is a Sequence
+            self._sequence[seq_start:seq_stop] = item
+        elif isinstance(index, numbers.Integral):
+            # Item is a symbol
+            self._sequence[index - self._seqstart] = item
+        else:
+            raise TypeError(
+                f"'{type(index).__name__}' instances are invalid indices"
+            )
+
+    def __eq__(self, item):
+        if not isinstance(item, AnnotatedSequence):
+            return False
+        return (    self.annotation == item.annotation
+                and self.sequence   == item.sequence
+                and self._seqstart  == item._seqstart)