Merge remote-tracking branch 'origin/staging'

docs/conf.py

@@ -68,9 +68,9 @@
 # built documents.
 #
 # The short X.Y version.
-version = '4.2'
+version = '4.3'
 # The full version, including alpha/beta/rc tags.
-release = '4.2.1'
+release = '4.3.0'
 
 
 # The language for content autogenerated by Sphinx. Refer to documentation

docs/courses.rst

@@ -0,0 +1,6 @@
+Courses
+=======
+
+- `Introductory pVACtools course <https://course.pvactools.org>`_
+- `Introductory immunotherapy workflow course <https://workflow-course.pvactools.org>`_
+- `Precision medicine course <https://pmbio.org/>`_

docs/funding.rst

@@ -0,0 +1,6 @@
+Funding
+=======
+
+pVACtools is supported by the National Cancer Institute (U01CA248235) and the
+V Foundation for Cancer Research (V2018-007) as well as a generous gift from
+Mrs. Cindy Goldberg and Mr. Evan Goldberg.

docs/index.rst

@@ -42,27 +42,60 @@ Contents
    :maxdepth: 1
 
    install
+   courses
    tools
    frequently_asked_questions
    releases
    license
    citation
+   funding
    contribute
    contact
    mailing_list
 
 New in Release |release|
 ------------------------
 
-This is a bugfix release. It fixes the following problem(s):
-
-- When running the reference protome similarity step with a reference protome
-  peptide fasta file and a species other than human or mouse, the run would be aborted
-  with an error that the refseq_protein_prot BLASTp database was incompatible with
-  the species. This error should not be emitted in this circumstance since
-  BLASTp is not run when using a reference proteome peptide fasta file. This
-  release fixes this error and allows users to run the reference proteome
-  similarity step on non-human and non-mouse data with a reference proteome peptide fasta.
+This is a minor feature release. It adds the following features:
+
+- Add a new helper command ``pvacseq|pvacfuse|pvacbind|pvacvector valid_algorithms``
+  by @ldhtnp in /~https://github.com/griffithlab/pVACtools/pull/1108
+- When running the ``pvacseq generate_protein_fasta`` command with the ``--phased-proximal-variants-vcf``
+  argument, output the intermediate ``proximal_variants.tsv`` file by @evelyn-schmidt
+  in /~https://github.com/griffithlab/pVACtools/pull/1091
+- In pVACview, clear the comment text input box after saving the comment by @ldhtnp
+  in /~https://github.com/griffithlab/pVACtools/pull/1113
+- Add support for mouse allele anchor positions by @ldhtnp in
+  /~https://github.com/griffithlab/pVACtools/pull/1110
+- Skip variants where VEP didn't predict an amino acid change by @susannasiebert
+  in /~https://github.com/griffithlab/pVACtools/pull/1121
+- Update the ordering of the fasta file output of the ``pvacseq|pvacfuse generate_protein_fasta``
+  command when running with the ``--input-tsv`` argument so that the order of the fasta sequences
+  is consistent with the order of the neoantigen candidates in the input TSV by @mhoang22 in
+  /~https://github.com/griffithlab/pVACtools/pull/1002
+- Updat the ``pvacfuse generate_protein_fasta`` command to allow aggregated TSVs as an input TSV
+  by @susannasiebert in /~https://github.com/griffithlab/pVACtools/pull/1134
+- Update pVACview to display the anchor positions currently applied to the data by @susannasiebert
+  in /~https://github.com/griffithlab/pVACtools/pull/1114
+
+This release also fixes the following bug(s):
+
+- Handle invalid pVACfuse characters by trimming the sequence instead of skipping it. The previous
+  implementation would lead to missing sequences in certain downstream steps, resulting in errors.
+  by @susannasiebert in /~https://github.com/griffithlab/pVACtools/pull/1130
+- Add new pVACview R files to the list of files getting copied into the pVACseq output folder.
+  These files were previsouly not copied in the the results folder, leading to error when running
+  the ``pvacview run`` commands on a pVACseq output directory. by @susannasiebert in
+  /~https://github.com/griffithlab/pVACtools/pull/1126
+- Remove single DP and DQ alpha and beta chain alleles from the list of supported alleles in MHCnuggetsII.
+  This is because those alleles need to be defined as a pair of alpha- and beta-chains in order to be
+  meaningful. Also remove DRA alleles from the same list since the DR locus is defined only by the beta
+  chain because functional variation in mature DRA gene products is absent. by @susannasiebert in
+  /~https://github.com/griffithlab/pVACtools/pull/1133
+- Fix errors in the rounding of the min and max values of the sliders in the custom pVACview module by
+  @evelyn-schmidt in /~https://github.com/griffithlab/pVACtools/pull/1116
+- Remove unused code in the Frameshift.pm VEP plugin that causes errors with certain types of variants
+  by @susannasiebert in /~https://github.com/griffithlab/pVACtools/pull/1122
 
 Past release notes can be found on our :ref:`releases` page.
 

docs/pvacbind/additional_commands.rst

@@ -19,6 +19,18 @@ List Valid Alleles
 
 .. program-output:: pvacbind valid_alleles -h
 
+.. _pvacbind_valid_algorithms:
+
+List Valid Algorithms
+---------------------
+
+.. program-output:: pvacbind valid_algorithms -h
+
+.. .. argparse::
+    :module: lib.valid_algorithms
+    :func: define_parser
+    :prog: pvacbind valid_algorithms
+
 List Allele-Specific Cutoffs
 ----------------------------
 

docs/pvacfuse/additional_commands.rst

@@ -29,6 +29,18 @@ List Valid Alleles
     :func: define_parser
     :prog: pvacfuse valid_alleles
 
+.. _pvacfuse_valid_algorithms:
+
+List Valid Algorithms
+---------------------
+
+.. program-output:: pvacfuse valid_algorithms -h
+
+.. .. argparse::
+    :module: lib.valid_algorithms
+    :func: define_parser
+    :prog: pvacfuse valid_algorithms
+
 List Allele-Specific Cutoffs
 ----------------------------
 

docs/pvacseq.rst

@@ -7,7 +7,11 @@
 pVACseq
 =======
 
-pVACseq is a cancer immunotherapy pipeline for the identification of **p**\ ersonalized **V**\ ariant **A**\ ntigens by **C**\ ancer **Seq**\ uencing (pVACseq) that integrates tumor mutation and expression data (DNA- and RNA-Seq). It enables cancer immunotherapy research by using massively parallel sequence data to predicting tumor-specific mutant peptides (neoantigens) that can elicit anti-tumor T cell immunity. It is being used in studies of checkpoint therapy response and to identify targets for personalized cancer vaccines and adoptive T cell therapies. For more general information, see the `manuscript published in Genome Medicine <http://www.genomemedicine.com/content/8/1/11>`_.
+pVACseq is a cancer immunotherapy pipeline for the identification of **p**\ ersonalized **V**\ ariant **A**\ ntigens by **C**\ ancer **Seq**\ uencing (pVACseq)
+that integrates tumor mutation and expression data (DNA- and RNA-Seq). It enables cancer immunotherapy research by using massively parallel sequence data to
+predict tumor-specific mutant peptides (neoantigens) that can elicit anti-tumor T cell immunity. It is being used in studies of checkpoint therapy response and
+to identify targets for personalized cancer vaccines and adoptive T cell therapies. For more general information, see the
+`manuscript published in Cancer Immunology Research <https://doi.org/10.1158/2326-6066.CIR-19-0401>`_.
 
 .. toctree::
    :maxdepth: 2

docs/pvacseq/additional_commands.rst

@@ -43,6 +43,18 @@ List Valid Alleles
     :func: define_parser
     :prog: pvacseq valid_alleles
 
+.. _valid_algorithms:
+
+List Valid Algorithms
+---------------------
+
+.. program-output:: pvacseq valid_algorithms -h
+
+.. .. argparse::
+    :module: lib.valid_algorithms
+    :func: define_parser
+    :prog: pvacseq valid_algorithms
+
 List Allele-Specific Cutoffs
 ----------------------------
 

docs/pvacvector/additional_commands.rst

@@ -41,6 +41,18 @@ List Valid Alleles
     :func: define_parser
     :prog: pvacfuse valid_alleles
 
+.. _pvacvector_valid_algorithms:
+
+List Valid Algorithms
+---------------------
+
+.. program-output:: pvacvector valid_algorithms -h
+
+.. .. argparse::
+    :module: lib.valid_algorithms
+    :func: define_parser
+    :prog: pvacvector valid_algorithms
+
 List Allele-Specific Cutoffs
 ----------------------------
 

docs/releases.rst

@@ -18,3 +18,4 @@ Release Notes
    releases/4_0
    releases/4_1
    releases/4_2
+   releases/4_3

docs/releases/4_3.rst

@@ -0,0 +1,46 @@
+Version 4.3
+===========
+
+Version 4.3.0
+-------------
+
+This is a minor feature release. It adds the following features:
+
+- Add a new helper command ``pvacseq|pvacfuse|pvacbind|pvacvector valid_algorithms``
+  by @ldhtnp in /~https://github.com/griffithlab/pVACtools/pull/1108
+- When running the ``pvacseq generate_protein_fasta`` command with the ``--phased-proximal-variants-vcf``
+  argument, output the intermediate ``proximal_variants.tsv`` file by @evelyn-schmidt
+  in /~https://github.com/griffithlab/pVACtools/pull/1091
+- In pVACview, clear the comment text input box after saving the comment by @ldhtnp
+  in /~https://github.com/griffithlab/pVACtools/pull/1113
+- Add support for mouse allele anchor positions by @ldhtnp in
+  /~https://github.com/griffithlab/pVACtools/pull/1110
+- Skip variants where VEP didn't predict an amino acid change by @susannasiebert
+  in /~https://github.com/griffithlab/pVACtools/pull/1121
+- Update the ordering of the fasta file output of the ``pvacseq|pvacfuse generate_protein_fasta``
+  command when running with the ``--input-tsv`` argument so that the order of the fasta sequences
+  is consistent with the order of the neoantigen candidates in the input TSV by @mhoang22 in
+  /~https://github.com/griffithlab/pVACtools/pull/1002
+- Updat the ``pvacfuse generate_protein_fasta`` command to allow aggregated TSVs as an input TSV
+  by @susannasiebert in /~https://github.com/griffithlab/pVACtools/pull/1134
+- Update pVACview to display the anchor positions currently applied to the data by @susannasiebert
+  in /~https://github.com/griffithlab/pVACtools/pull/1114
+
+This release also fixes the following bug(s):
+
+- Handle invalid pVACfuse characters by trimming the sequence instead of skipping it. The previous
+  implementation would lead to missing sequences in certain downstream steps, resulting in errors.
+  by @susannasiebert in /~https://github.com/griffithlab/pVACtools/pull/1130
+- Add new pVACview R files to the list of files getting copied into the pVACseq output folder.
+  These files were previsouly not copied in the the results folder, leading to error when running
+  the ``pvacview run`` commands on a pVACseq output directory. by @susannasiebert in
+  /~https://github.com/griffithlab/pVACtools/pull/1126
+- Remove single DP and DQ alpha and beta chain alleles from the list of supported alleles in MHCnuggetsII.
+  This is because those alleles need to be defined as a pair of alpha- and beta-chains in order to be
+  meaningful. Also remove DRA alleles from the same list since the DR locus is defined only by the beta
+  chain because functional variation in mature DRA gene products is absent. by @susannasiebert in
+  /~https://github.com/griffithlab/pVACtools/pull/1133
+- Fix errors in the rounding of the min and max values of the sliders in the custom pVACview module by
+  @evelyn-schmidt in /~https://github.com/griffithlab/pVACtools/pull/1116
+- Remove unused code in the Frameshift.pm VEP plugin that causes errors with certain types of variants
+  by @susannasiebert in /~https://github.com/griffithlab/pVACtools/pull/1122

pvactools/lib/__init__.py

@@ -11,6 +11,7 @@
     "fasta_generator",
     "output_parser",
     "valid_alleles",
+    'valid_algorithms',
     'net_chop',
     "netmhc_stab",
     "filter",

pvactools/lib/aggregate_all_epitopes.py

@@ -8,6 +8,9 @@
 from abc import ABCMeta, abstractmethod
 import itertools
 import csv
+import glob
+import ast
+from pvactools.lib.run_utils import get_anchor_positions
 
 from pvactools.lib.prediction_class import PredictionClass
 
@@ -123,6 +126,17 @@ def determine_used_prediction_algorithms(self):
                 prediction_algorithms.append(algorithm)
         return prediction_algorithms
 
+    def determine_used_epitope_lengths(self):
+        col_name = self.determine_epitope_seq_column_name()
+        return list(set([len(s) for s in pd.read_csv(self.input_file, delimiter="\t", usecols=[col_name])[col_name]]))
+
+    def determine_epitope_seq_column_name(self):
+        headers = pd.read_csv(self.input_file, delimiter="\t", nrows=0).columns.tolist()
+        for header in ["MT Epitope Seq", "Epitope Seq"]:
+            if header in headers:
+                return header
+        raise Exception("No mutant epitope sequence header found.")
+
     def problematic_positions_exist(self):
         headers = pd.read_csv(self.input_file, delimiter="\t", nrows=0).columns.tolist()
         return 'Problematic Positions' in headers
@@ -160,18 +174,18 @@ def determine_columns_used_for_aggregation(self, prediction_algorithms, el_algor
 
     def set_column_types(self, prediction_algorithms):
         dtypes = {
-            'Chromosome': str,
+            'Chromosome': "string",
             "Start": "int32",
             "Stop": "int32",
-            'Reference': str,
-            'Variant': str,
+            'Reference': "string",
+            'Variant': "string",
             "Variant Type": "category",
             "Mutation Position": "category",
             "Median MT IC50 Score": "float32",
             "Median MT Percentile": "float32",
             "Best MT IC50 Score": "float32",
             "Best MT Percentile": "float32",
-            "Protein Position": "str",
+            "Protein Position": "string",
             "Transcript Length": "int32",
         }
         for algorithm in prediction_algorithms:
@@ -183,6 +197,7 @@ def set_column_types(self, prediction_algorithms):
 
     def execute(self):
         prediction_algorithms = self.determine_used_prediction_algorithms()
+        epitope_lengths = self.determine_used_epitope_lengths()
         el_algorithms = self.determine_used_el_algorithms()
         used_columns = self.determine_columns_used_for_aggregation(prediction_algorithms, el_algorithms)
         dtypes = self.set_column_types(prediction_algorithms)
@@ -209,6 +224,7 @@ def execute(self):
                 'allele_specific_anchors': self.allele_specific_anchors,
                 'alleles': self.hla_types.tolist(),
                 'anchor_contribution_threshold': self.anchor_contribution_threshold,
+                'epitope_lengths': epitope_lengths,
             }
         else:
             metrics = {}
@@ -281,6 +297,20 @@ def __init__(
                     probs[hla] = line
             anchor_probabilities[length] = probs
         self.anchor_probabilities = anchor_probabilities
+
+        mouse_anchor_positions = {}
+        for length in [8, 9, 10, 11]:
+            base_dir = os.path.abspath(os.path.join(os.path.dirname(os.path.realpath(__file__)), '..'))
+            file_name = os.path.join(base_dir, 'tools', 'pvacview', 'data', "mouse_anchor_predictions_{}_mer.tsv".format(length))
+            values = {}
+            with open(file_name, 'r') as fh:
+                reader = csv.DictReader(fh, delimiter="\t")
+                for line in reader:
+                    allele = line.pop('Allele')
+                    values[allele] = {int(k): ast.literal_eval(v) for k, v in line.items()}
+            mouse_anchor_positions[length] = values
+        self.mouse_anchor_positions = mouse_anchor_positions
+
         self.allele_specific_anchors = allele_specific_anchors
         self.anchor_contribution_threshold = anchor_contribution_threshold
         super().__init__()
@@ -362,7 +392,7 @@ def is_anchor_residue_pass(self, mutation):
             binding_threshold = self.binding_threshold
 
         anchor_residue_pass = True
-        anchors = self.get_anchor_positions(mutation['HLA Allele'], len(mutation['MT Epitope Seq']))
+        anchors = get_anchor_positions(mutation['HLA Allele'], len(mutation['MT Epitope Seq']), self.allele_specific_anchors, self.anchor_probabilities, self.anchor_contribution_threshold, self.mouse_anchor_positions)
         # parse out mutation position from str
         position = mutation["Mutation Position"]
         if pd.isna(position):
@@ -382,20 +412,6 @@ def is_anchor_residue_pass(self, mutation):
                     anchor_residue_pass = False
         return anchor_residue_pass
 
-    def get_anchor_positions(self, hla_allele, epitope_length):
-        if self.allele_specific_anchors and epitope_length in self.anchor_probabilities and hla_allele in self.anchor_probabilities[epitope_length]:
-            probs = self.anchor_probabilities[epitope_length][hla_allele]
-            positions = []
-            total_prob = 0
-            for (pos, prob) in sorted(probs.items(), key=lambda x: x[1], reverse=True):
-                total_prob += float(prob)
-                positions.append(int(pos))
-                if total_prob > self.anchor_contribution_threshold:
-                    return positions
-
-        return [1, 2, epitope_length - 1 , epitope_length]
-
-
     #assign mutations to a "Classification" based on their favorability
     def get_tier(self, mutation, vaf_clonal):
         if self.use_allele_specific_binding_thresholds and mutation['HLA Allele'] in self.allele_specific_binding_thresholds:
@@ -731,10 +747,11 @@ def sort_table(self, df):
     def copy_pvacview_r_files(self):
         module_dir = os.path.dirname(__file__)
         r_folder = os.path.abspath(os.path.join(module_dir,"..","tools","pvacview"))
+        files = glob.iglob(os.path.join(r_folder, "*.R"))
         destination = os.path.abspath(os.path.dirname(self.output_file))
         os.makedirs(os.path.join(destination, "www"), exist_ok=True)
-        for i in ["ui.R", "app.R", "server.R", "styling.R", "anchor_and_helper_functions.R"]:
-            shutil.copy(os.path.join(r_folder, i), os.path.join(destination, i))
+        for i in files:
+            shutil.copy(i, destination)
         for i in ["anchor.jpg", "pVACview_logo.png", "pVACview_logo_mini.png"]:
             shutil.copy(os.path.join(r_folder, "www", i), os.path.join(destination, "www", i))