Limiting dgidb query results to specific sources (#55)

* Add parameter for dgidb data sources

* Added test

* Update src/cli.yaml

Co-authored-by: Johannes Köster <johannes.koester@tu-dortmund.de>

Co-authored-by: Johannes Köster <johannes.koester@tu-dortmund.de>

Author: FelixMoelder

src/bcf/annotate_dgidb.rs

@@ -33,18 +33,28 @@ pub fn annotate_dgidb(
     vcf_path: &str,
     api_path: String,
     field_name: &str,
+    datasources: Option<Vec<&str>>,
     genes_per_request: usize,
 ) -> Result<(), Box<dyn Error>> {
-    let gene_drug_interactions = request_interaction_drugs(vcf_path, api_path, genes_per_request)?;
+    let gene_drug_interactions =
+        request_interaction_drugs(vcf_path, api_path, datasources, genes_per_request)?;
     modify_vcf_entries(vcf_path, gene_drug_interactions, field_name)
 }
 
 fn request_interaction_drugs(
     vcf_path: &str,
     api_path: String,
+    datasources_opt: Option<Vec<&str>>,
     genes_per_request: usize,
 ) -> Result<Option<HashMap<String, Vec<(String, Vec<String>)>>>, Box<dyn Error>> {
     let mut genes = collect_genes(vcf_path)?;
+    let datasources = if let Some(entries) = datasources_opt {
+        let mut b = String::from("&interaction_sources=");
+        b.push_str(entries.join(",").as_str());
+        b
+    } else {
+        String::new()
+    };
     if genes.is_empty() {
         return Ok(None);
     }
@@ -57,6 +67,7 @@ fn request_interaction_drugs(
     {
         let mut slice_api_path = api_path.clone();
         slice_api_path.push_str(gene_slice.join(",").as_str());
+        slice_api_path.push_str(datasources.as_str());
         let res: Dgidb = reqwest::get(&slice_api_path)?.json()?;
 
         for term in res.matched_terms {

src/cli.yaml

@@ -187,6 +187,14 @@ subcommands:
               short: f
               default_value: dgiDB_drugs
               help: Info field name to be used for annotation.
+          - datasources:
+              long: sources
+              short: s
+              value_name: str
+              multiple: true
+              help: |
+                A list of data sources included in query. If omitted all sources are considered.
+                A list of all sources can be found at http://dgidb.org/api/v2/interaction_sources.json
           - genes-per-request:
               long: genes-per-request
               short: g

src/main.rs

@@ -65,6 +65,10 @@ fn main() -> Result<(), Box<dyn Error>> {
             &matches.value_of("vcf").unwrap(),
             matches.value_of("api-path").unwrap().to_string(),
             &matches.value_of("field").unwrap(),
+            match matches.is_present("datasources") {
+                true => Some(matches.values_of("datasources").unwrap().collect()),
+                false => None,
+            },
             value_t!(matches, "genes-per-request", usize).unwrap(),
         ),
         ("oncoprint", Some(matches)) => {

tests/expected/annotate_dgidb_drugbank_test.bcf

@@ -249,6 +249,19 @@ fn test_vcf_annotate_dgidb() {
     );
 }
 
+#[test]
+fn test_vcf_annotate_dgidb_drugbank() {
+    assert!(
+        Command::new("bash")
+            .arg("-c")
+            .arg("target/debug/rbt vcf-annotate-dgidb tests/annotate_dgidb_test.vcf -s DrugBank> /tmp/annotate_dgidb_drugbank_test.bcf")
+            .spawn().unwrap().wait().unwrap().success());
+    test_output(
+        "/tmp/annotate_dgidb_drugbank_test.bcf",
+        "tests/expected/annotate_dgidb_drugbank_test.bcf",
+    );
+}
+
 #[test]
 fn test_stats_fasta_file() {
     assert!(Command::new("bash")