wip: blastp

backend/Dockerfile

@@ -1,24 +1,29 @@
 FROM quay.io/rosalindfranklininstitute/blast:2.16.0 AS blast
 
-FROM python:3.12-bookworm AS builder
+FROM python:3.13-bookworm AS builder
 
-RUN pip install --user wheel pipenv
+RUN pip install --upgrade pip wheel pipenv
+
+RUN adduser --uid 10191 --group --system --no-create-home nonroot
+RUN chown -R nonroot:nonroot /usr/src
+USER nonroot
+WORKDIR /usr/src
 
 # Tell pipenv to create venv in the current directory
 ENV PIPENV_VENV_IN_PROJECT=1
 
 COPY Pipfile Pipfile.lock pyproject.toml setup.cfg /usr/src/
-WORKDIR /usr/src
 
-RUN /root/.local/bin/pipenv sync
+ARG PIPENV_SYNC_FLAGS=
+RUN pipenv sync ${PIPENV_SYNC_FLAGS}
 
 COPY manage.py /usr/src/
 COPY antigendjango /usr/src/antigendjango
 COPY antigenapi /usr/src/antigenapi
 
 RUN DJANGO_CI=true .venv/bin/python manage.py collectstatic --noinput
 
-FROM python:3.12-slim-bookworm AS prod
+FROM python:3.13-slim-bookworm AS prod
 
 # liblmdb-dev required by BLAST
 RUN apt-get update && apt-get install -y \
@@ -27,7 +32,10 @@ RUN apt-get update && apt-get install -y \
     liblmdb-dev \
     && rm -rf /var/lib/apt/lists/*
 
-RUN mkdir -v /usr/src/.venv
+RUN adduser --uid 10191 --group --system --no-create-home nonroot
+RUN chown -R nonroot:nonroot /usr/src
+USER nonroot
+WORKDIR /usr/src
 
 # blastp and makeblastdb commands
 COPY --from=blast /blast/ReleaseMT/bin/blastp /usr/local/bin/blastp
@@ -42,23 +50,17 @@ COPY uwsgi.ini /usr/src
 
 WORKDIR /usr/src
 
-RUN addgroup --gid 10191 nonroot
-RUN adduser --uid 10191 --gid 10191 --system --no-create-home nonroot
-USER nonroot
-
 CMD [".venv/bin/uwsgi", "--ini", "uwsgi.ini"]
 
 FROM builder AS dev
 
 # liblmdb-dev required by BLAST
+USER root
 RUN apt update && apt install -y liblmdb-dev && rm -rf /var/lib/apt/lists/*
+USER nonroot
 
 # blastp and makeblastdb commands
 COPY --from=blast /blast/ReleaseMT/bin/blastp /usr/local/bin/blastp
 COPY --from=blast /blast/ReleaseMT/bin/makeblastdb /usr/local/bin/makeblastdb
 
-ENV PATH="$PATH:/root/.local/bin:/usr/src/.venv/bin"
-
-RUN pipenv sync --dev
-
 CMD ["pipenv", "run", "python", "manage.py", "runserver", "0.0.0.0:8080"]

backend/antigenapi/bioinformatics/blast.py

@@ -0,0 +1,137 @@
+import os
+import subprocess
+from tempfile import TemporaryDirectory
+from typing import Optional
+
+from ..models import SequencingRunResults
+from .imgt import read_airr_file
+
+# https://www.ncbi.nlm.nih.gov/books/NBK279684/table/appendices.T.options_common_to_all_blast/
+BLAST_FMT_MULTIPLE_FILE_BLAST_JSON = "15"
+BLAST_NUM_THREADS = 4
+
+
+def get_db_fasta(include_run: Optional[int] = None, exclude_run: Optional[int] = None):
+    """Get the sequencing database in fasta format.
+
+    Args:
+        include_run (int, optional): Sequencing run ID to include.
+          Defaults to None.
+        exclude_run (int, optional): Sequencing run ID to exclude.
+          Defaults to None.
+
+    Returns:
+        str: Sequencing run as a FASTA format string
+    """
+    fasta_data = ""
+    query = SequencingRunResults.objects.all()
+    if include_run:
+        query = query.filter(sequencing_run_id=include_run)
+    if exclude_run:
+        query = query.exclude(sequencing_run_id=exclude_run)
+    for sr in query:
+        airr_file = read_airr_file(
+            sr.airr_file, usecols=("sequence_id", "sequence_alignment_aa")
+        )
+        airr_file = airr_file[airr_file.sequence_alignment_aa.notna()]
+        if not airr_file.empty:
+            for _, row in airr_file.iterrows():
+                fasta_data += f"> {row.sequence_id}\n"
+                fasta_data += f"{row.sequence_alignment_aa.replace('.', '')}\n"
+    return fasta_data
+
+
+def get_sequencing_run_fasta(sequencing_run_id: int):
+    """Get sequencing run in BLAST format.
+
+    Args:
+        sequencing_run_id (int): Sequencing run ID
+
+    Returns:
+        str: Sequencing run as a FASTA format string
+    """
+    return get_db_fasta(include_run=sequencing_run_id)
+
+
+def run_blastp(
+    sequencing_run_id: int, outfmt: str = BLAST_FMT_MULTIPLE_FILE_BLAST_JSON
+):
+    """Run blastp for a sequencing run vs rest of database.
+
+    Args:
+        sequencing_run_id (int): Sequencing run ID.
+
+    Returns:
+        JSONResponse: Single file BLAST JSON
+    """
+    db_data = get_db_fasta()
+    if not db_data:
+        return None
+    query_data = get_sequencing_run_fasta(sequencing_run_id)
+    if not query_data:
+        return None
+
+    # Write the DB to disk as .fasta format
+    with TemporaryDirectory() as tmp_dir:
+        fasta_filename = os.path.join(tmp_dir, "db.fasta")
+        with open(fasta_filename, "w") as f:
+            f.write(db_data)
+
+        # Run makeblastdb
+        mkdb_proc = subprocess.run(
+            [
+                "makeblastdb",
+                "-in",
+                "db.fasta",
+                "-dbtype",
+                "prot",
+                "-out",
+                "antigen.db",
+            ],
+            capture_output=True,
+            cwd=tmp_dir,
+        )
+
+        if mkdb_proc.returncode != 0:
+            raise Exception(
+                f"makeblastdb returned exit code of "
+                f"{mkdb_proc.returncode}\n\n"
+                f"STDOUT: {mkdb_proc.stdout}\n\n"
+                f"STDERR: {mkdb_proc.stderr}"
+            )
+
+        # Write out query file
+        fasta_filename = os.path.join(tmp_dir, "query.fasta")
+        with open(fasta_filename, "w") as f:
+            f.write(query_data)
+
+        # Run blastp
+        blastp_proc = subprocess.run(
+            [
+                "blastp",
+                "-db",
+                "antigen.db",
+                "-query",
+                "query.fasta",
+                "-outfmt",
+                outfmt,
+                "-out",
+                "antigen.results",
+                "-num_threads",
+                str(BLAST_NUM_THREADS),
+            ],
+            capture_output=True,
+            cwd=tmp_dir,
+        )
+
+        if blastp_proc.returncode != 0:
+            raise Exception(
+                f"blastp returned exit code of "
+                f"{blastp_proc.returncode}\n\n"
+                f"STDOUT: {blastp_proc.stdout}\n\n"
+                f"STDERR: {blastp_proc.stderr}"
+            )
+
+        # Read in the results file
+        with open(os.path.join(tmp_dir, "antigen.results"), "r") as f:
+            return f.read()

backend/antigenapi/models.py

@@ -28,7 +28,7 @@
 )
 from django.db.models.signals import post_init, post_save
 
-from .bioinformatics import read_airr_file
+from .bioinformatics.imgt import read_airr_file
 
 
 class Project(Model):

backend/antigenapi/tests/test_parsers.py

@@ -2,7 +2,7 @@
 import unittest
 import zipfile
 
-from antigenapi.bioinformatics import load_sequences, trim_sequence
+from antigenapi.bioinformatics.imgt import load_sequences, trim_sequence
 from antigenapi.views_old import _extract_well
 
 

backend/antigenapi/urls.py

@@ -4,7 +4,6 @@
 from antigenapi.views.dashboard import AuditLogLatestEvents, DashboardStats
 from antigenapi.views_old import (
     AntigenViewSet,
-    BlastAllView,
     CohortViewSet,
     ElisaPlateViewSet,
     GlobalFastaView,
@@ -30,5 +29,4 @@
     path("fasta/", GlobalFastaView.as_view(), name="fasta"),
     path("dashboard/stats", DashboardStats.as_view(), name="dashboard_stats"),
     path("dashboard/latest", AuditLogLatestEvents.as_view(), name="dashboard_latest"),
-    path("blast/all", BlastAllView.as_view(), name="blast_all"),
 ]