doc updates and more TSV output

Author: sigven

R/main.R

@@ -260,6 +260,25 @@ write_cpsr_output <- function(report,
         ".classification.tsv.gz"
       )
     )
+  fnames[["tsv_bm"]] <-
+    file.path(
+      output_dir,
+      paste0(
+        sample_fname_pattern,
+        ".biomarker_evidence.tsv.gz"
+      )
+    )
+
+  fnames[["tsv_pgx"]] <-
+    file.path(
+      output_dir,
+      paste0(
+        sample_fname_pattern,
+        ".pgx_findings.tsv.gz"
+      )
+    )
+
+
   fnames[["xlsx"]] <-
     file.path(
       output_dir,
@@ -372,12 +391,11 @@ write_cpsr_output <- function(report,
 
   if (output_format == "tsv") {
     if (NROW(
-      report[["content"]][["snv_indel"]]$callset$tsv
-    ) > 0) {
+      report[["content"]][["snv_indel"]]$callset$tsv) > 0) {
       pcgrr::log4r_info("------")
       pcgrr::log4r_info(
         paste0(
-          "Generating SNV/InDel tab-separated values file (.tsv.gz) ",
+          "Generating tab-separated values file (.tsv.gz) ",
           "with variant findings"
         )
       )
@@ -389,6 +407,57 @@ write_cpsr_output <- function(report,
         na = "."
       )
     }
+    ## Biomarker TSV
+    if (NROW(report$content$snv_indel$callset$variant$bm) > 0) {
+      biomarker_tsv <- report$content$snv_indel$callset$variant$bm |>
+        dplyr::mutate(
+          BM_MOLECULAR_PROFILE = pcgrr::strip_html(
+            .data$BM_MOLECULAR_PROFILE
+          )
+        ) |>
+        dplyr::select(
+          dplyr::any_of(
+            cpsr::col_format_output[["xlsx_biomarker"]]
+          ))
+
+      pcgrr::log4r_info("------")
+      pcgrr::log4r_info(
+        paste0(
+          "Generating tab-separated values file (.tsv.gz) ",
+          "with biomarker evidence"
+        )
+      )
+      readr::write_tsv(
+        biomarker_tsv,
+        file = fnames[["tsv_bm"]],
+        col_names = T,
+        quote = "none",
+        na = "."
+      )
+    }
+
+    if (NROW(report[["content"]]$snv_indel$callset$variant$pgx) > 0) {
+      pgx_tsv <- report[["content"]]$snv_indel$callset$variant$pgx |>
+        dplyr::select(
+          dplyr::any_of(
+            cpsr::col_format_output[["xlsx_pgx"]]
+          ))
+      pcgrr::log4r_info("------")
+      pcgrr::log4r_info(
+        paste0(
+          "Generating tab-separated values file (.tsv.gz) ",
+          "with pharmacogenomic evidence"
+        )
+      )
+      readr::write_tsv(
+        pgx_tsv,
+        file = fnames[["tsv_pgx"]],
+        col_names = T,
+        quote = "none",
+        na = "."
+      )
+    }
+
   }
 
   if (output_format == "xlsx") {

inst/templates/quarto/cpsr_gwas.qmd

@@ -70,8 +70,8 @@ DT::datatable(variants_gwas_cancer,
     "GENOTYPE",
     backgroundColor = 
       DT::styleEqual(
-        "hom_ref",
-        "grey"
+        cpsr::color_palette[['genotypes']][['levels']],
+        cpsr::color_palette[['genotypes']][['values']]
       )
   )
 

vignettes/CHANGELOG.Rmd

@@ -5,7 +5,7 @@ output: rmarkdown::html_document
 
 ## v2.2.0
 
-- Date: **2025-03-XX**
+- Date: **2025-03-22**
 - Major data updates
   - ClinVar (2025-03)
   - dbNSFP (v5.0)
@@ -14,7 +14,7 @@ output: rmarkdown::html_document
   - PanelApp (2025-02)
   - UniProt KB (2025-01)
   - Cancer Gene Census (v101)
-- Added more cancer susceptibility genes in panel zero (ATG12, BIK, CHD1L, CMTR2, 
+- Added more cancer susceptibility genes to panel zero (ATG12, BIK, CHD1L, CMTR2, 
 CPAP, HAVCR2, LLGL2, MYO3A, MYO5B, PAH, TTC7A)
 - Added a pharmacogenetic findings option (`--pgx_findings`), which will include pharmacogenetic findings in the HTML report (within the `Genomic biomarkers` section)
    - For now, this is implemented very simple, not considering star alleles, but merely focusing on pathogenic variants or drug-response related variants in DPYD, TPMT, and NUDT15

vignettes/installation.Rmd

@@ -4,7 +4,11 @@ output: rmarkdown::html_document
 ---
 
 CPSR is distributed alongside the [Personal Cancer Genome Reporter (PCGR)](https://github.com/sigven/pcgr), so please follow
-the [PCGR installation steps](https://sigven.github.io/pcgr/articles/installation.html) to install CPSR, either through [Docker](https://docs.docker.com/), [Apptainer/Singularity](https://apptainer.org/docs/user/latest/index.html), or [Conda](https://docs.conda.io/projects/conda/en/latest/user-guide/getting-started.html).
+the [PCGR installation steps](https://sigven.github.io/pcgr/articles/installation.html) to install CPSR, either through 
+
+- [Docker](https://sigven.github.io/pcgr/articles/installation.html#b--docker), or
+- [Apptainer/Singularity](https://sigven.github.io/pcgr/articles/installation.html#c--singularityapptainer), or 
+- [Conda](https://sigven.github.io/pcgr/articles/installation.html#a--conda).
 
 We recommend Conda as the simplest framework to install PCGR and CPSR, using either a MacOS or a Linux platform.
 

vignettes/output.Rmd

@@ -45,16 +45,17 @@ The report is structured in multiple sections, described briefly below:
 	    * Annotation resources
 	       * Information on annotation sources utilized by CPSR, including versions and licensing requirements
 	    * Variant classification
-	       * Overview of how CPSR performs variant classification of variants not recorded in ClinVar, listing ACMG criteria and associated scores
+	       * Overview of how CPSR performs variant annotation and classification of variants not recorded in ClinVar, 
+	         listing ACMG criteria and associated scores, calibration of classification thresholds etc.
 
   8. __References__
-	    * Supporting scientific literature - knowledge resources, guideline references etc.)
+	    * Supporting scientific literature - knowledge resources, guideline references etc.
 
 <br><br>
 
 ### Variant call format - VCF
 
-A VCF file containing annotated, germline calls (single nucleotide variants and insertions/deletions) is generated with the following naming convention:
+A VCF file containing annotated, germline variant calls (single nucleotide variants and insertions/deletions) is generated with the following naming convention:
 
 - `<sample_id>.cpsr.<genome_assembly>.vcf.gz (.tbi)`
   - The __sample_id__ is provided as input by the user, and reflects a unique identifier of the sample to be analyzed. Following common standards, the annotated VCF file is compressed with [bgzip](http://www.htslib.org/doc/bgzip.html) and indexed with [tabix](http://www.htslib.org/doc/tabix.html). Below follows a description of all annotations/tags present in the VCF INFO column after processing with the CPSR annotation pipeline:
@@ -322,6 +323,8 @@ be present if any data is found):
 
 ### Tab-separated values - TSV
 
+#### _Variant classification_
+
 We provide a compressed tab-separated values file with variant classifications and the most essential variant/gene annotations. The file has the following naming convention:
 
 - `<sample_id>.cpsr.<genome_assembly>.classification.tsv.gz`
@@ -403,6 +406,19 @@ The following variables are included in the tiered TSV file (VCF tags in the que
 
 **NOTE**: The user has the possibility to append the TSV file with data from other INFO tags in the input VCF (i.e. using the *--retained_info_tags* option)
 
+#### _Biomarker evidence_
+
+We provide a compressed tab-separated values file with variants implicated as germline biomarkers. The file has the following naming convention:
+
+- `<sample_id>.cpsr.<genome_assembly>.biomarker_evidence.tsv.gz`
+
+#### _Pharmacogenetic findings_
+
+We provide a compressed tab-separated values file with variants implicated with drug toxicity/dosage effects of cancer chemotherapies. The file has the following naming convention:
+
+- `<sample_id>.cpsr.<genome_assembly>.pgx_findings.tsv.gz`
+
+
 <br><br>
 
 ### Biomarker annotations

vignettes/running.Rmd

@@ -19,7 +19,7 @@ The user can flexibly choose the set of cancer predisposition genes for which in
 
 The user can choose from a range of pre-defined gene panels, selected from the following list of panel identifiers:
 
-- **0**:  [Exploratory panel - all cancer predisposition gene](https://sigven.github.io/cpsr/articles/virtual_panels.html)
+- **0**:  [Exploratory panel - all cancer predisposition genes](https://sigven.github.io/cpsr/articles/virtual_panels.html)
 - **1**:  [Adult solid tumours cancer susceptibility](https://panelapp.genomicsengland.co.uk/panels/245/)
 - **2**:  [Adult solid tumours for rare disease](https://panelapp.genomicsengland.co.uk/panels/391/)
 - **3**:  [Bladder cancer pertinent cancer susceptibility](https://panelapp.genomicsengland.co.uk/panels/208/)
@@ -65,10 +65,11 @@ The user can choose from a range of pre-defined gene panels, selected from the f
 - **43**:  [Upper gastrointestinal cancer pertinent cancer susceptibility](https://panelapp.genomicsengland.co.uk/panels/273/)
 - **44**:  [DNA repair genes pertinent cancer susceptibility](https://panelapp.genomicsengland.co.uk/panels/256/)
 
+<br>
 
 #### Custom-made virtual gene panels
 
-CPSR allows users to create custom virtual gene panels for reporting.  Any set of genes found in the [CPSR superpanel (panel 0)](virtual_panels.html#panel-0) can be used to design a custom virtual gene panel. Technically, the users need to create a simple one-column text file with Ensembl gene identifiers, and provide a name for the custom panel:
+CPSR allows users to create custom virtual gene panels for reporting.  Any set of genes found in the [CPSR superpanel (panel 0)](virtual_panels.html#panel-0) can be used to design a custom virtual gene panel. Technically, the users need to create a simple one-column text (TSV) file with Ensembl gene identifiers, and provide a name for the custom panel, using the following command line options:
 
 * `--custom_list <custom_list_tsv>`
 * `--custom_list_name <custom_list_name`
@@ -104,6 +105,8 @@ By default, CPSR do not report variants in the input sample that are found in ca
 
 * `--clinvar_report_noncancer`
 
+<br>
+
 ### Optional report contents
 
 CPSR allows users to report recommended [incidental findings](https://www.ncbi.nlm.nih.gov/clinvar/docs/acmg/), the occurrence of important variants with respec to chemotherapy toxicity, and also the genotypes of reported cancer risk loci from [genome-wide association studies (GWAS)](https://www.ebi.ac.uk/gwas/):
@@ -144,7 +147,7 @@ Panel options:
   --panel_id VIRTUAL_PANEL_ID
                         Comma-separated string with identifier(s) of predefined virtual cancer predisposition gene panels,
                         choose any combination of the following identifiers (GEP = Genomics England PanelApp):
-                        0 = CPSR exploratory cancer predisposition panel (PanelApp genes / TCGA's germline study / Cancer Gene Census / Other)
+                        0 = CPSR exploratory cancer predisposition panel (PanelApp genes / TCGA's germline study / Cancer Gene Census / Other )
                         1 = Adult solid tumours cancer susceptibility (GEP)
                         2 = Adult solid tumours for rare disease (GEP)
                         3 = Bladder cancer pertinent cancer susceptibility (GEP)
@@ -249,14 +252,15 @@ Report generation with the example VCF, using the [Adult solid tumours cancer su
 $ (base) conda activate pcgr
 $ (pcgr)
 cpsr \
-	 --input_vcf ~/cpsr-2.0.0/inst/examples/example.vcf.gz \
+	 --input_vcf ~/cpsr-2.2.0/inst/examples/example.vcf.gz \
 	 --vep_dir ~/.vep \
 	 --refdata_dir ~/pcgr_ref_data \
-	 --output_dir ~/cpsr-2.0.0/ \
+	 --output_dir ~/cpsr-2.2.0/ \
 	 --genome_assembly grch37 \
 	 --panel_id 1 \
 	 --sample_id example \
 	 --secondary_findings \
+	 --pgx_findings \
 	 --classify_all \
 	 --maf_upper_threshold 0.2 \
 	 --force_overwrite
@@ -273,7 +277,9 @@ This command will produce the following output files in the _output_ folder:
   5. __example.cpsr.grch37.xlsx__ - An Excel workbook that contains
       * _i)_ information on virtual gene panel interrogated for variants
       * _ii)_ classification of clinical significance for variants overlapping with cancer predisposition genes
-      * _iii)_ match of variants with existing biomarkers (if any found)
-      * _iv)_ secondary findings (if any found)
+      * _iii)_ secondary findings (if any found)
+      * _iv)_ match of variants with existing biomarkers (if any found)
+      * _v)_ overlap with pharmacogenomic variants (if any found)
+     
   6. __example.cpsr.grch37.html__ - Interactive HTML report with clinically relevant variants in cancer predisposition genes
   7. __example.cpsr.grch37.snvs_indels.classification.tsv.gz__ - TSV file with key annotations of germline SNVs/InDels classified according to clinical significance

vignettes/variant_classification.Rmd

@@ -37,7 +37,7 @@ while(i <= NROW(cpsr::acmg$evidence_codes)){
 ```
 
 <br>
-Currently (as of 2025-02), based on a calibration against ClinVar-classified variants (minimum two review status stars) in n = 105 core cancer predisposition genes, the clinical significance (**CPSR_CLASSIFICATION**) is determined based on the following ranges of pathogenicity scores:
+Currently (as of March 2025), based on a calibration against ClinVar-classified variants (minimum two review status stars) in n = 105 core cancer predisposition genes, the clinical significance (**CPSR_CLASSIFICATION**) is determined based on the following ranges of pathogenicity scores:
 
 <br>
 

vignettes/virtual_panels.Rmd

@@ -7,14 +7,14 @@ output: rmarkdown::html_document
 
 The cancer predisposition report can show variants found in a number of well-known cancer predisposition genes, and the specific set of genes can be customized by the user by choosing any of the following __virtual gene panels (0 - 44)__:
 
-  * **Panel 0** is a non-conservative, research-based _superpanel_ assembled through multiple sources on cancer predisposition genes:
-	* A list of 152 genes that were curated and established within TCGA’s pan-cancer study ([Huang et al., *Cell*, 2018](https://www.ncbi.nlm.nih.gov/pubmed/29625052))
-	* A list of 114 protein-coding genes that has been manually curated in COSMIC’s [Cancer Gene Census v100](https://cancer.sanger.ac.uk/census),
+* **Panel 0** is a non-conservative, research-based _superpanel_ assembled through multiple sources on cancer predisposition genes:
+	* A list of 151 genes that were curated and established within TCGA’s pan-cancer study ([Huang et al., *Cell*, 2018](https://www.ncbi.nlm.nih.gov/pubmed/29625052))
+	* A list of 114 protein-coding genes that has been manually curated in COSMIC’s [Cancer Gene Census v101](https://cancer.sanger.ac.uk/census),
 	* Genes from all [Genomics England PanelApp](https://panelapp.genomicsengland.co.uk/) panels for inherited cancers and tumor syndromes, as well as DNA repair genes (detailed below)
-	* Additional genes deemed relevant for cancer predisposition (i.e. contributed by CPSR users)
+	* Additional genes deemed relevant for cancer predisposition (i.e. contributed by CPSR users etc.)
 
 
-	The combination of the above sources resulted in a non-redundant set of **n = 572**
+	The combination of the above sources resulted in a non-redundant set of **n = 574**
 	genes of relevance for cancer predisposition (see complete details [below](#panel-0))
 
 	Data with respect to mechanisms of inheritance (<i>MoI</i> - autosomal recessive (AR) vs. autosomal