ENH: add ProfileHMM[*] semantic types

ci/recipe/meta.yaml

@@ -26,6 +26,7 @@ requirements:
     - h5py
     - qiime2 {{ qiime2_epoch }}.*
     - samtools
+    - pyhmmer
 
 test:
   commands:

q2_types/__init__.py

@@ -31,3 +31,4 @@
 importlib.import_module('q2_types.genome_data')
 importlib.import_module('q2_types.kaiju')
 importlib.import_module('q2_types.reference_db')
+importlib.import_module('q2_types.profile_hmms')

q2_types/profile_hmms/__init__.py

@@ -0,0 +1,38 @@
+# ----------------------------------------------------------------------------
+# Copyright (c) 2023, QIIME 2 development team.
+#
+# Distributed under the terms of the Modified BSD License.
+#
+# The full license is in the file LICENSE, distributed with this software.
+# ----------------------------------------------------------------------------
+from ._format import (
+    PressedProfileHmmsDirectoryFmt,
+    ProteinSingleProfileHmmDirectoryFmt,
+    ProteinMultipleProfileHmmDirectoryFmt,
+    DNASingleProfileHmmDirectoryFmt,
+    DNAMultipleProfileHmmDirectoryFmt,
+    RNASingleProfileHmmDirectoryFmt,
+    RNAMultipleProfileHmmDirectoryFmt
+)
+from ._type import (
+    ProfileHMM,
+    SingleProtein, SingleDNA, SingleRNA,
+    MultipleProtein, MultipleDNA, MultipleRNA,
+    PressedRNA, PressedDNA, PressedProtein
+)
+
+__all__ = [
+    "ProfileHmmBinaryFileFmt",
+    "PressedProfileHmmsDirectoryFmt",
+    "ProfileHmmFileFmt",
+    "ProteinSingleProfileHmmDirectoryFmt",
+    "ProteinMultipleProfileHmmDirectoryFmt",
+    "DNASingleProfileHmmDirectoryFmt",
+    "DNAMultipleProfileHmmDirectoryFmt",
+    "RNASingleProfileHmmDirectoryFmt",
+    "RNAMultipleProfileHmmDirectoryFmt",
+    "ProfileHMM",
+    "SingleProtein", "SingleDNA", "SingleRNA",
+    "MultipleProtein", "MultipleDNA", "MultipleRNA",
+    "PressedRNA", "PressedDNA", "PressedProtein"
+]

q2_types/profile_hmms/_format.py

@@ -0,0 +1,138 @@
+# ----------------------------------------------------------------------------
+# Copyright (c) 2023, QIIME 2 development team.
+#
+# Distributed under the terms of the Modified BSD License.
+#
+# The full license is in the file LICENSE, distributed with this software.
+# ----------------------------------------------------------------------------
+from pyhmmer.plan7 import HMMFile
+from pyhmmer.errors import AlphabetMismatch
+from qiime2.plugin import model
+from qiime2.core.exceptions import ValidationError
+from q2_types.plugin_setup import plugin
+
+
+class ProfileHmmBinaryFileFmt(model.BinaryFileFormat):
+    def _validate_(self, level):
+        pass
+
+
+class PressedProfileHmmsDirectoryFmt(model.DirectoryFormat):
+    """
+    The  <hmmfile>.h3m file contains the profile HMMs
+    and their annotation in a binary format. The <hmmfile>.h3i file is an
+    SSI index for the <hmmfile>.h3m file.  The <hmmfile>.h3f file contains
+    precomputed data structures for the fast heuristic filter
+    (the MSV filter).  The <hmmfile>.h3p file contains precomputed data
+    structures for the rest of each profile.
+    """
+    h3m = model.File(r'.*\.hmm\.h3m', format=ProfileHmmBinaryFileFmt)
+    h3i = model.File(r'.*\.hmm\.h3i', format=ProfileHmmBinaryFileFmt)
+    h3f = model.File(r'.*\.hmm\.h3f', format=ProfileHmmBinaryFileFmt)
+    h3p = model.File(r'.*\.hmm\.h3p', format=ProfileHmmBinaryFileFmt)
+
+
+class ProfileHmmFileFmt(model.TextFileFormat):
+    def _validate_(self, level: str):
+        """
+        Check http://eddylab.org/software/hmmer/Userguide.pdf
+        section "HMMER profile HMM files" for full description of
+        hmm file format.
+        """
+        parse_n_profiles = {"min": 3, "max": None}[level]
+        tolerance = 0.0001
+
+        with HMMFile(str(self)) as hmm_file:
+            try:
+                hmm_profiles = list(hmm_file)
+            except AlphabetMismatch:
+                raise ValidationError(
+                    "Found profiles with alphabet different from "
+                    f"'{self.alphabet}'"
+                )
+
+            if len(hmm_profiles) > 1 and self.single:
+                raise ValidationError(
+                        f"Expected 1 profile, found {len(hmm_profiles)}."
+                    )
+
+            for hmm_profile in hmm_profiles[:parse_n_profiles]:
+                hmm_profile.validate(tolerance=tolerance)
+
+                if hmm_profile.alphabet.type.lower() != self.alphabet:
+                    raise ValidationError(
+                        "Found profile with alphabet "
+                        f"{hmm_profile.alphabet.type.lower()}\n"
+                        f"Expected alphabet: {self.alphabet}."
+                    )
+
+
+class ProteinProfileHmmFileFmt(ProfileHmmFileFmt):
+    alphabet = "amino"
+
+
+class ProteinSingleProfileHmmFileFmt(ProteinProfileHmmFileFmt):
+    single = True
+
+
+class ProteinMultipleProfileHmmFileFmt(ProteinProfileHmmFileFmt):
+    single = False
+
+
+class DNAProfileHmmFileFmt(ProfileHmmFileFmt):
+    alphabet = "dna"
+
+
+class DNASingleProfileHmmFileFmt(DNAProfileHmmFileFmt):
+    single = True
+
+
+class DNAMultipleProfileHmmFileFmt(DNAProfileHmmFileFmt):
+    single = False
+
+
+class RNAProfileHmmFileFmt(ProfileHmmFileFmt):
+    alphabet = "rna"
+
+
+class RNASingleProfileHmmFileFmt(RNAProfileHmmFileFmt):
+    single = True
+
+
+class RNAMultipleProfileHmmFileFmt(RNAProfileHmmFileFmt):
+    single = False
+
+
+class ProteinSingleProfileHmmDirectoryFmt(model.DirectoryFormat):
+    profile = model.File(r'.*\.hmm', format=ProteinSingleProfileHmmFileFmt)
+
+
+class ProteinMultipleProfileHmmDirectoryFmt(model.DirectoryFormat):
+    profiles = model.File(r'.*\.hmm', format=ProteinMultipleProfileHmmFileFmt)
+
+
+class DNASingleProfileHmmDirectoryFmt(model.DirectoryFormat):
+    profile = model.File(r'.*\.hmm', format=DNASingleProfileHmmFileFmt)
+
+
+class DNAMultipleProfileHmmDirectoryFmt(model.DirectoryFormat):
+    profiles = model.File(r'.*\.hmm', format=DNAMultipleProfileHmmFileFmt)
+
+
+class RNASingleProfileHmmDirectoryFmt(model.DirectoryFormat):
+    profile = model.File(r'.*\.hmm', format=RNASingleProfileHmmFileFmt)
+
+
+class RNAMultipleProfileHmmDirectoryFmt(model.DirectoryFormat):
+    profiles = model.File(r'.*\.hmm', format=RNAMultipleProfileHmmFileFmt)
+
+
+plugin.register_formats(
+    PressedProfileHmmsDirectoryFmt,
+    ProteinSingleProfileHmmDirectoryFmt,
+    ProteinMultipleProfileHmmDirectoryFmt,
+    DNASingleProfileHmmDirectoryFmt,
+    DNAMultipleProfileHmmDirectoryFmt,
+    RNASingleProfileHmmDirectoryFmt,
+    RNAMultipleProfileHmmDirectoryFmt
+)

q2_types/profile_hmms/_type.py

@@ -0,0 +1,136 @@
+# ----------------------------------------------------------------------------
+# Copyright (c) 2023, QIIME 2 development team.
+#
+# Distributed under the terms of the Modified BSD License.
+#
+# The full license is in the file LICENSE, distributed with this software.
+# ----------------------------------------------------------------------------
+from qiime2.plugin import SemanticType
+from q2_types.plugin_setup import plugin
+from q2_types.profile_hmms._format import (
+    PressedProfileHmmsDirectoryFmt,
+    DNASingleProfileHmmDirectoryFmt,
+    DNAMultipleProfileHmmDirectoryFmt,
+    RNASingleProfileHmmDirectoryFmt,
+    RNAMultipleProfileHmmDirectoryFmt,
+    ProteinSingleProfileHmmDirectoryFmt,
+    ProteinMultipleProfileHmmDirectoryFmt
+)
+
+
+ProfileHMM = SemanticType('ProfileHMM', field_names='type')
+SingleProtein = SemanticType(
+    'SingleProtein', variant_of=ProfileHMM.field['type']
+)
+SingleDNA = SemanticType(
+    'SingleDNA', variant_of=ProfileHMM.field['type']
+)
+SingleRNA = SemanticType(
+    'SingleRNA', variant_of=ProfileHMM.field['type']
+)
+MultipleProtein = SemanticType(
+    'MultipleProtein', variant_of=ProfileHMM.field['type']
+)
+MultipleDNA = SemanticType(
+    'MultipleDNA', variant_of=ProfileHMM.field['type']
+)
+MultipleRNA = SemanticType(
+    'MultipleRNA', variant_of=ProfileHMM.field['type']
+)
+PressedProtein = SemanticType(
+    'PressedProtein', variant_of=ProfileHMM.field['type']
+)
+PressedDNA = SemanticType(
+    'PressedDNA', variant_of=ProfileHMM.field['type']
+)
+PressedRNA = SemanticType(
+    'PressedRNA', variant_of=ProfileHMM.field['type']
+)
+
+plugin.register_semantic_types(
+    ProfileHMM,
+    SingleProtein, SingleDNA, SingleRNA,
+    MultipleProtein, MultipleDNA, MultipleRNA,
+    PressedProtein, PressedDNA, PressedRNA
+)
+
+plugin.register_artifact_class(
+    ProfileHMM[PressedProtein],
+    directory_format=PressedProfileHmmsDirectoryFmt,
+    description=(
+        "A collection of profile Hidden Markov Models for amino acid "
+        "sequences in binary format and indexed."
+    )
+)
+
+plugin.register_artifact_class(
+    ProfileHMM[PressedDNA],
+    directory_format=PressedProfileHmmsDirectoryFmt,
+    description=(
+        "A collection of profile Hidden Markov Models for DNA "
+        "sequences in binary format and indexed."
+    )
+)
+
+plugin.register_artifact_class(
+    ProfileHMM[PressedRNA],
+    directory_format=PressedProfileHmmsDirectoryFmt,
+    description=(
+        "A collection of profile Hidden Markov Models for RNA "
+        "sequences in binary format and indexed."
+    )
+)
+
+plugin.register_artifact_class(
+    ProfileHMM[SingleProtein],
+    directory_format=ProteinSingleProfileHmmDirectoryFmt,
+    description=(
+        "One single profile Hidden Markov Model representing a group "
+        "of related proteins."
+    )
+)
+
+plugin.register_artifact_class(
+    ProfileHMM[SingleDNA],
+    directory_format=DNASingleProfileHmmDirectoryFmt,
+    description=(
+        "One single profile Hidden Markov Model representing a group "
+        "of related DNA sequences."
+    )
+)
+
+plugin.register_artifact_class(
+    ProfileHMM[SingleRNA],
+    directory_format=RNASingleProfileHmmDirectoryFmt,
+    description=(
+        "One single profile Hidden Markov Model representing a group "
+        "of related RNA sequences."
+    )
+)
+
+plugin.register_artifact_class(
+    ProfileHMM[MultipleProtein],
+    directory_format=ProteinMultipleProfileHmmDirectoryFmt,
+    description=(
+        "A collection of profile Hidden Markov Models, "
+        "each representing a group of related proteins."
+    )
+)
+
+plugin.register_artifact_class(
+    ProfileHMM[MultipleDNA],
+    directory_format=DNAMultipleProfileHmmDirectoryFmt,
+    description=(
+        "A collection of profile Hidden Markov Models, "
+        "each representing a group of related DNA sequences."
+    )
+)
+
+plugin.register_artifact_class(
+    ProfileHMM[MultipleRNA],
+    directory_format=RNAMultipleProfileHmmDirectoryFmt,
+    description=(
+        "A collection of profile Hidden Markov Models, "
+        "each representing a group of related RNA sequences."
+    )
+)

q2_types/profile_hmms/tests/__init__.py

@@ -0,0 +1,7 @@
+# ----------------------------------------------------------------------------
+# Copyright (c) 2023, QIIME 2 development team.
+#
+# Distributed under the terms of the Modified BSD License.
+#
+# The full license is in the file LICENSE, distributed with this software.
+# ----------------------------------------------------------------------------