feat: adding param --partial-matching-lower-bound to pandora map, dis…

…cover and command subcommands

include/compare_main.h

@@ -60,6 +60,7 @@ struct CompareOptions {
     float min_allele_fraction_covg_gt { 0 };
     float genotyping_error_rate { 0.01 };
     uint16_t confidence_threshold { 1 };
+    float partial_matching_lower_bound { 0.5 };
     bool keep_extra_debugging_files { false };
 };
 

include/denovo_discovery/discover_main.h

@@ -39,6 +39,7 @@ struct DiscoverOptions {
     bool no_gene_coverage_filtering { false };
     uint32_t min_cluster_size { 10 };
     uint32_t max_num_kmers_to_avg { 100 };
+    float partial_matching_lower_bound { 0.5 };
     bool keep_extra_debugging_files { false };
 };
 

include/map_main.h

@@ -62,6 +62,7 @@ struct MapOptions {
     float min_allele_fraction_covg_gt { 0 };
     float genotyping_error_rate { 0.01 };
     uint16_t confidence_threshold { 1 };
+    float partial_matching_lower_bound { 0.5 };
     bool keep_extra_debugging_files { false };
 };
 

include/utils.h

@@ -115,7 +115,7 @@ uint32_t pangraph_from_read_file(const SampleData& sample,
     const float conflicting_clusters_overlap_threshold=0.8,
     const float conflicting_clusters_minimiser_tolerance=0.05,
     uint32_t threads = 1, const bool keep_extra_debugging_files = false,
-    const uint32_t rng_seed = 0);
+    const uint32_t rng_seed = 0, const float partial_matching_lower_bound=0.5);
 
 void infer_most_likely_prg_path_for_pannode(
     const std::vector<std::shared_ptr<LocalPRG>>&, PanNode*, uint32_t, float);

src/compare_main.cpp

@@ -198,6 +198,17 @@ void setup_compare_subcommand(CLI::App& app)
         ->type_name("FLOAT")
         ->group("Mapping");
 
+    description
+        = "Allows for partial matching between reads and a PRG. If this value is for e.g. 0.5, it means that "
+          "pandora will match a read to a PRG if the cluster of hits has size at least 0.5 * expected cluster size "
+          "for the given error rate and kmer value. Lower values allow for more hits, but possibly for false positive "
+          "matches.";
+    compare_subcmd
+        ->add_option("--partial-matching-lower-bound", opt->partial_matching_lower_bound, description)
+        ->capture_default_str()
+        ->type_name("FLOAT")
+        ->group("Mapping");
+
     description = "Maximum number of kmers to average over when selecting the maximum "
                   "likelihood path";
     compare_subcmd->add_option("--kmer-avg", opt->max_num_kmers_to_avg, description)
@@ -326,7 +337,8 @@ int pandora_compare(CompareOptions& opt)
             opt.max_diff, opt.error_rate, sample_outdir,
             opt.min_cluster_size, opt.genome_size, opt.max_covg,
             opt.conflicting_clusters_overlap_threshold, opt.conflicting_clusters_minimiser_tolerance,
-            opt.threads, opt.keep_extra_debugging_files, opt.rng_seed);
+            opt.threads, opt.keep_extra_debugging_files, opt.rng_seed,
+            opt.partial_matching_lower_bound);
 
         if (pangraph_sample->nodes.empty()) {
             BOOST_LOG_TRIVIAL(warning)

src/denovo_discovery/discover_main.cpp

@@ -75,6 +75,17 @@ void setup_discover_subcommand(CLI::App& app)
         ->type_name("FLOAT")
         ->group("Mapping");
 
+    description
+        = "Allows for partial matching between reads and a PRG. If this value is for e.g. 0.5, it means that "
+          "pandora will match a read to a PRG if the cluster of hits has size at least 0.5 * expected cluster size "
+          "for the given error rate and kmer value. Lower values allow for more hits, but possibly for false positive "
+          "matches.";
+    discover_subcmd
+        ->add_option("--partial-matching-lower-bound", opt->partial_matching_lower_bound, description)
+        ->capture_default_str()
+        ->type_name("FLOAT")
+        ->group("Mapping");
+
     description
         = "When two clusters of hits are conflicting, the one with highest number of unique minimisers "
           "will be kept. However, if the difference between the number of unique minimisers is too small, "
@@ -241,7 +252,8 @@ void pandora_discover_core(const SampleData& sample, Index &index, DiscoverOptio
         = pangraph_from_read_file(sample, pangraph, index, opt.max_diff, opt.error_rate, sample_outdir,
         opt.min_cluster_size, opt.genome_size, opt.max_covg,
         opt.conflicting_clusters_overlap_threshold, opt.conflicting_clusters_minimiser_tolerance,
-        opt.threads, opt.keep_extra_debugging_files, opt.rng_seed);
+        opt.threads, opt.keep_extra_debugging_files, opt.rng_seed,
+        opt.partial_matching_lower_bound);
 
     if (pangraph->nodes.empty()) {
         BOOST_LOG_TRIVIAL(warning)

src/map_main.cpp

@@ -87,6 +87,17 @@ void setup_map_subcommand(CLI::App& app)
         ->type_name("FLOAT")
         ->group("Mapping");
 
+    description
+        = "Allows for partial matching between reads and a PRG. If this value is for e.g. 0.5, it means that "
+          "pandora will match a read to a PRG if the cluster of hits has size at least 0.5 * expected cluster size "
+          "for the given error rate and kmer value. Lower values allow for more hits, but possibly for false positive "
+          "matches.";
+    map_subcmd
+        ->add_option("--partial-matching-lower-bound", opt->partial_matching_lower_bound, description)
+        ->capture_default_str()
+        ->type_name("FLOAT")
+        ->group("Mapping");
+
     map_subcmd
         ->add_flag("--kg", opt->output_kg,
             "Save kmer graphs with forward and reverse coverage annotations for found "
@@ -322,7 +333,8 @@ int pandora_map(MapOptions& opt)
         = pangraph_from_read_file(sample, pangraph, index, opt.max_diff, opt.error_rate,
             opt.outdir, opt.min_cluster_size, opt.genome_size, opt.max_covg,
             opt.conflicting_clusters_overlap_threshold, opt.conflicting_clusters_minimiser_tolerance,
-            opt.threads, opt.keep_extra_debugging_files, opt.rng_seed);
+            opt.threads, opt.keep_extra_debugging_files, opt.rng_seed,
+            opt.partial_matching_lower_bound);
 
     if (pangraph->nodes.empty()) {
         BOOST_LOG_TRIVIAL(info) << "Found none of the LocalPRGs in the reads.";

src/utils.cpp

@@ -544,15 +544,16 @@ uint32_t pangraph_from_read_file(const SampleData& sample,
     const float conflicting_clusters_overlap_threshold,
     const float conflicting_clusters_minimiser_tolerance,
     uint32_t threads,
-    const bool keep_extra_debugging_files, const uint32_t rng_seed)
+    const bool keep_extra_debugging_files, const uint32_t rng_seed,
+    const float partial_matching_lower_bound)
 {
     // constant variables
     const SampleIdText sample_name = sample.first;
     const SampleFpath sample_filepath = sample.second;
     const std::string tag = "[Sample " + sample_name + "]: ";
     const uint32_t w = index.get_window_size();
     const uint32_t k = index.get_kmer_size();
-    const double fraction_kmers_required_for_cluster = 0.1 / exp(e_rate * k);
+    const double fraction_kmers_required_for_cluster = partial_matching_lower_bound / exp(e_rate * k);
     const uint32_t nb_reads_to_map_in_a_batch = 1000;
 
     BOOST_LOG_TRIVIAL(trace) << tag << "e_rate: " << e_rate;