Merge pull request #831 from Clinical-Genomics/release_8.2.4

feat: Balsamic Release 8.2.4

BALSAMIC/assets/scripts/create_pdf.py

@@ -0,0 +1,105 @@
+#!/usr/bin/env python
+import os
+import click
+
+from fpdf import FPDF
+from PIL import Image
+
+
+@click.command(
+    short_help="Merge images and a txt file into a single PDF",
+)
+@click.argument("output", type=click.Path(exists=False), required=True)
+@click.argument("data", type=click.Path(exists=True), required=True)
+@click.argument("images", nargs=-1, type=click.Path(exists=True), required=True)
+def create_pdf(output, data, images):
+    pdf = generate_fpdf()
+    pdf = add_table_pdf(pdf, data)
+    pdf = add_images_pdf(pdf, images)
+    pdf.output(output)
+
+
+class PDF(FPDF):
+    def footer(self):
+        self.set_y(-15)
+        self.set_font("helvetica", "I", 8)
+        self.cell(0, 10, f"Page {self.page_no()}/{{nb}}", 0, 0, "C")
+
+
+def generate_fpdf():
+    pdf = PDF()
+    pdf.alias_nb_pages(alias="{nb}")
+    return pdf
+
+
+def add_images_pdf(pdf, img_paths):
+    pdf.set_font("helvetica", "B", 15)
+
+    for path in img_paths:
+        title = os.path.basename(path).replace(".png", "")
+
+        # Image & page layout parameters
+        if "sunrise" in title:
+            page_orientation = "portrait"
+            img_size = 500, 500
+            title_w_pos = 25
+            title_wh = 140, 10
+            img_xy = 10, 55
+        else:
+            page_orientation = "landscape"
+            img_size = 800, 800
+            title_w_pos = 68.5
+            title_wh = 140, 10
+            img_xy = 5, 40
+
+        pdf.add_page(orientation=page_orientation)
+
+        # Title position & styling
+        pdf.cell(title_w_pos)
+        pdf.cell(title_wh[0], title_wh[1], title, 1, 0, "C")
+
+        # Image position & resizing
+        img = Image.open(path)
+        img.thumbnail(img_size, Image.ANTIALIAS)
+        pdf.image(img, img_xy[0], img_xy[1])
+
+    return pdf
+
+
+def add_table_pdf(pdf, data_path):
+
+    with open(data_path) as data:
+        data = data.readlines()
+
+    pdf.add_page()
+    pdf.set_font("helvetica", "B", 15)
+
+    # Title layout & styling
+    title = os.path.basename(data_path).replace(".txt", "")
+    pdf.cell(25)
+    pdf.cell(140, 10, title, 1, 0, "C")
+    pdf.cell(35, 25, ln=1)  # Post title indentation
+
+    # Table layout & styling
+    pdf.set_font("Times", size=11)
+    line_height = pdf.font_size * 2.5
+    col_width = pdf.epw / 4  # Even distribution of the content
+    for row in data:
+        pdf.cell(45)
+        for statistic in row.split():
+            pdf.multi_cell(
+                col_width,
+                line_height,
+                statistic,
+                align="C",
+                border=1,
+                ln=3,
+                max_line_height=pdf.font_size,
+            )
+        pdf.ln(line_height)
+
+    return pdf
+
+
+if __name__ == "__main__":
+    create_pdf()

BALSAMIC/config/cluster.json

@@ -251,5 +251,9 @@
 	"ascat_tumor_normal": {
 		"time": "8:00:00",
 		"n": 36
+	},
+	"ascat_tumor_normal_merge_output": {
+		"time": "00:15:00",
+		"n": 1
 	}
 }

BALSAMIC/constants/workflow_params.py

@@ -118,7 +118,7 @@
         "column_info": "-c 1 -S 2 -E 3 -g 4",
     },
     "vep": {
-        "vep_filters": "--compress_output bgzip --vcf --everything --allow_non_variant --dont_skip --buffer_size 20000 --format vcf --offline --variant_class --merged --cache --verbose --force_overwrite"
+        "vep_filters": "--compress_output bgzip --vcf --everything --hgvsg --allow_non_variant --dont_skip --buffer_size 20000 --format vcf --offline --variant_class --merged --cache --verbose --force_overwrite"
     },
     "umicommon": {
         "align_header": "'@RG\\tID:{sample}\\tSM:{sample}\\tLB:TargetPanel\\tPL:ILLUMINA'",

BALSAMIC/constants/workflow_rules.py

@@ -136,5 +136,5 @@
     "mergeBam_normal",
     "cnvkit_paired",
     "cnvkit_single",
-    "ascat_tumor_normal",
+    "ascat_tumor_normal_merge_output",
 ]

BALSAMIC/snakemake_rules/variant_calling/somatic_sv_tumor_normal.rule

@@ -125,21 +125,18 @@ rule ascat_tumor_normal:
         gccorrection = config["reference"]["ascat_gccorrection"],
         chryloci= config["reference"]["ascat_chryloci"],
     output:
-        finalvcf = temp(vcf_dir + "CNV.somatic." + config["analysis"]["case_id"] + ".ascat.vcf.gz"),
-        samplestatistics = vcf_dir + "CNV.somatic." + config["analysis"]["case_id"] + ".ascat.samplestatistics.txt",
-        germlineplot = vcf_dir + "CNV.somatic." + config["analysis"]["case_id"] + ".ascat.germline.png",
-        rawprofileplot = vcf_dir + "CNV.somatic." + config["analysis"]["case_id"] + ".ascat.rawprofile.png",
-        sunriseplot = vcf_dir + "CNV.somatic." + config["analysis"]["case_id"] + ".ascat.sunrise.png",
-        tumorplot = vcf_dir + "CNV.somatic." + config["analysis"]["case_id"] + ".ascat.tumor.png",
-        aspcfplot = vcf_dir + "CNV.somatic." + config["analysis"]["case_id"] + ".ascat.ASPCF.png",
-        ascatprofileplot = vcf_dir + "CNV.somatic." + config["analysis"]["case_id"] + ".ascat.ascatprofile.png",
-        namemap = temp(vcf_dir + "CNV.somatic." + config["analysis"]["case_id"] + ".ascat.sample_name_map"),
+        final_vcf = vcf_dir + "CNV.somatic." + config["analysis"]["case_id"] + ".ascat.vcf.gz",
+        sample_statistics = vcf_dir + "CNV.somatic." + config["analysis"]["case_id"] + ".ascat.samplestatistics.txt",
+        ascat_plots= expand(
+            vcf_dir + "CNV.somatic." + config["analysis"]["case_id"] + ".ascat." + "{output_suffix}" + ".png",
+            output_suffix=["ascatprofile", "rawprofile", "ASPCF", "tumor", "germline", "sunrise"]
+        ),
+        namemap = vcf_dir + "CNV.somatic." + config["analysis"]["case_id"] + ".ascat.sample_name_map",
     benchmark:
         benchmark_dir + 'ascat_tumor_normal_' + config["analysis"]["case_id"] + "_ascat.tsv"
     singularity:
         Path(singularity_image, config["bioinfo_tools"].get("ascatNgs") + ".sif").as_posix()
     params:
-        housekeeper_id = {"id": config["analysis"]["case_id"], "tags": "ascat-ngs"},
         tmpdir = tempfile.mkdtemp(prefix=tmp_dir),
         tumor = "TUMOR",
         normal = "NORMAL",
@@ -165,25 +162,46 @@ ascat.pl \
   -tumour {input.bamT} \
   -normal {input.bamN};
 
-cp  {params.tmpdir}/{params.tumor}.copynumber.caveman.vcf.gz {output.finalvcf};
+cp  {params.tmpdir}/{params.tumor}.copynumber.caveman.vcf.gz {output.final_vcf};
 
-cp {params.tmpdir}/{params.tumor}.samplestatistics.txt {output.samplestatistics};
+cp {params.tmpdir}/{params.tumor}.samplestatistics.txt {output.sample_statistics};
 
-cp {params.tmpdir}/{params.tumor}.germline.png {output.germlineplot};
+cp {params.tmpdir}/{params.tumor}.ASCATprofile.png {output.ascat_plots[0]};
 
-cp {params.tmpdir}/{params.tumor}.rawprofile.png {output.rawprofileplot};
+cp {params.tmpdir}/{params.tumor}.rawprofile.png {output.ascat_plots[1]};
 
-cp {params.tmpdir}/{params.tumor}.sunrise.png {output.sunriseplot};
+cp {params.tmpdir}/{params.tumor}.ASPCF.png {output.ascat_plots[2]};
 
-cp {params.tmpdir}/{params.tumor}.tumour.png {output.tumorplot};
+cp {params.tmpdir}/{params.tumor}.tumour.png {output.ascat_plots[3]};
 
-cp {params.tmpdir}/{params.tumor}.ASCATprofile.png {output.ascatprofileplot};
+cp {params.tmpdir}/{params.tumor}.germline.png {output.ascat_plots[4]};
 
-cp {params.tmpdir}/{params.tumor}.ASPCF.png {output.aspcfplot};
+cp {params.tmpdir}/{params.tumor}.sunrise.png {output.ascat_plots[5]};
 
-tabix -p vcf -f {output.finalvcf};
+tabix -p vcf -f {output.final_vcf};
 
 echo -e \"{params.tumor}\\tTUMOR\\n{params.normal}\\tNORMAL\" > {output.namemap};
 
 rm -rf {params.tmpdir};
         """
+
+rule ascat_tumor_normal_merge_output:
+    input:
+        sample_statistics = vcf_dir + "CNV.somatic." + config["analysis"]["case_id"] + ".ascat.samplestatistics.txt",
+        ascat_plots= expand(
+            vcf_dir + "CNV.somatic." + config["analysis"]["case_id"] + ".ascat." + "{output_suffix}" + ".png",
+            output_suffix=["ascatprofile", "rawprofile", "ASPCF", "tumor", "germline", "sunrise"]
+        ),
+    output:
+        ascat_output_pdf = vcf_dir + "CNV.somatic." + config["analysis"]["case_id"] + ".ascat.output.pdf"
+    params:
+        housekeeper_id = {"id": config["analysis"]["case_id"], "tags": "research"},
+        merge_ascat_output_script= get_script_path("create_pdf.py"),
+    threads:
+        get_threads(cluster_config, "ascat_tumor_normal_merge_output")
+    message:
+        "Merge the ascatNgs output plots together with the sample statistics into a single PDF"
+    shell:
+        """
+        python {params.merge_ascat_output_script} {output.ascat_output_pdf} {input.sample_statistics} {input.ascat_plots}
+        """

BALSAMIC/workflows/balsamic.smk

@@ -241,6 +241,9 @@ if config["analysis"]["sequencing_type"] == "wgs" and config['analysis']['analys
         analysis_specific_results.extend([Path(result_dir, "dragen", "SNV.somatic." + config["analysis"]["case_id"] + ".dragen_tumor.bam").as_posix(),
                                           Path(result_dir, "dragen", "SNV.somatic." + config["analysis"]["case_id"] + ".dragen.vcf.gz").as_posix()])
 
+if config["analysis"]["sequencing_type"] == "wgs" and config['analysis']['analysis_type'] == "paired":
+    analysis_specific_results.append(expand(vcf_dir + "{vcf}.output.pdf", vcf=get_vcf(config, ["ascat"], [config["analysis"]["case_id"]])))
+
 if 'benchmark_plots' in config:
     log_dir = config["analysis"]["log"]
     if not check_executable("sh5util"):

CHANGELOG.rst

@@ -1,3 +1,12 @@
+[8.2.4]
+-------
+
+Added:
+^^^^^^
+
+* ``--hgvsg`` annotation to VEP #830
+* ``ascatNgs`` PDF delivery (plots & statistics) #828
+
 [8.2.3]
 -------
 Fixed:

setup.py

@@ -19,13 +19,15 @@
         "colorclass>=2.2.0",
         "coloredlogs>=14.0",
         "cyvcf2<0.10.0",
+        "fpdf2>=2.4.6",
         "graphviz==0.16",
         "gsutil==4.50",
         "jinja2>=2.11.2",
         "matplotlib>=3.3.0",
         "networkx>=2.4",
         "numpy>=1.19.2",
         "pandas>1.1.0",
+        "pillow>=8.4.0",
         "psutil>=5.7.0",
         "pydantic>=1.5.1",
         "pygments>=2.6.1",

tests/scripts/test_create_pdf.py

@@ -0,0 +1,74 @@
+from pathlib import Path
+
+from fpdf import FPDF
+
+from BALSAMIC.assets.scripts.create_pdf import (
+    generate_fpdf,
+    add_images_pdf,
+    add_table_pdf,
+    create_pdf,
+)
+
+
+def test_generate_fpdf():
+    # WHEN creating a dummy FPDF file
+    pdf = generate_fpdf()
+
+    # THEN check if the pdf has been correctly created
+    assert isinstance(pdf, FPDF)
+
+
+def test_add_images_pdf():
+    # GIVEN ascatNGgs output PNG images
+    test_images_path = [
+        "tests/test_data/ascat_output/CNV.somatic.SAMPLE.ascat.sunrise.png",
+        "tests/test_data/ascat_output/CNV.somatic.SAMPLE.ascat.germline.png",
+    ]
+
+    # WHEN calling the function
+    pdf = add_images_pdf(generate_fpdf(), test_images_path)
+
+    # THEN check if the images are appended to the PDF
+    assert isinstance(pdf, FPDF)
+    assert pdf.page_no() == 2
+
+
+def test_add_table_pdf():
+    # GIVEN ascatNGgs output sample statistics .txt
+    test_statistics_path = (
+        "tests/test_data/ascat_output/CNV.somatic.SAMPLE.ascat.samplestatistics.txt"
+    )
+
+    # WHEN calling the function
+    pdf = add_table_pdf(generate_fpdf(), test_statistics_path)
+
+    # THEN check if the table is appended to the created PDF
+    assert isinstance(pdf, FPDF)
+    assert pdf.page_no() == 1
+
+
+def test_create_pdf(tmp_path, cli_runner):
+    # GIVEN ascatNGgs output statistics
+    statistics_path = (
+        "tests/test_data/ascat_output/CNV.somatic.SAMPLE.ascat.samplestatistics.txt"
+    )
+
+    # GIVEN ascatNGgs output plots
+    plots_path = [
+        "tests/test_data/ascat_output/CNV.somatic.SAMPLE.ascat.germline.png",
+        "tests/test_data/ascat_output/CNV.somatic.SAMPLE.ascat.sunrise.png",
+    ]
+
+    # GIVEN the output path
+    output_path = tmp_path / "ascat.output.pdf"
+
+    print(output_path)
+
+    # WHEN invoking the python script
+    result = cli_runner.invoke(
+        create_pdf, [str(output_path), statistics_path, plots_path[0], plots_path[1]]
+    )
+
+    # THEN check if the PDF is correctly created and there is no errors
+    assert result.exit_code == 0
+    assert Path(output_path).exists()

tests/test_data/ascat_output/CNV.somatic.SAMPLE.ascat.samplestatistics.txt

@@ -0,0 +1,7 @@
+NormalContamination 0.378222936036507
+Ploidy 2.69008904657384
+rho 0.55
+psi 2.75
+goodnessOfFit 93.9311185291303
+GenderChr Y
+GenderChrFound N