fixing validation calcs

bin/combine-summary.R

@@ -23,6 +23,7 @@ col.list <- c("ID",
          "TAXA",
          "SEGMENT",
          "REFERENCE",
+         "ASSEMBLY_VARIANT",
          "ASSEMBLY_NC",
          "ASSEMBLY_QC",
          "ASSEMBLY_QC_REASON",
@@ -84,13 +85,14 @@ df <- df %>%
          ASSEMBLY_QC_REASON = gsub(ASSEMBLY_QC_REASON, pattern = '; $', replacement = ''))
 # summarize assembly variants
 df <- df %>%
-  group_by(TAXA, SEGMENT) %>%
-  mutate(ASSEMBLY_VARIANT = paste0(row_number()," / ", n()))
+  group_by(ID, TAXA, SEGMENT) %>%
+  mutate(ASSEMBLY_VARIANT = paste0(row_number()," of ", n())) %>%
+  ungroup()
 
 # order columns
 df <- df %>%
   drop_na(ID) %>%
   select(col.list) %>%
   select(-ASSEMBLY_NC)
 # save combined summary
-write.csv(x=df, file="combined-summary.csv", quote = F, row.names = F)
+write.csv(x=df, file="combined-summary.csv", row.names = F)

bin/ref-select_accurate.R

@@ -27,7 +27,8 @@ paf <- read_tsv(paf_file, col_names = F) %>%
          LENGTH = 7,
          START = 8,
          END = 9,
-         ALIGN = 11) %>%
+         ALIGN = 11,
+         QUAL = 12) %>%
   mutate(ASSEMBLY=str_remove_all(TARGET, pattern="(\\d+$)"),
          CONTIG = str_extract(TARGET, "(\\d+$)")) %>%
   select(TARGET, ASSEMBLY, CONTIG, LENGTH, START, END, ALIGN)

bin/val_gather.sh

@@ -39,15 +39,15 @@ S2=$(cat $ALN | grep '>' | sed -n 2p | tr -d '>\n')
 # Second sequence (S2) in the alignment is treated as the "sample"
 # True Positives (TP) = all sites in S2 that match S1
 # True Negative (TN) = does not apply
-# False Positives (FP) = sites in S2 that are absent in S1 (i.e., "A", "T", "C", or "G" in S2 and "-" in S1)
+# False Positives (FP) = sites in S2 that are absent in S1 (i.e., "A", "T", "C", or "G" in S2 and "-" in S1) or sites in S2 that differ from S1, excluding absent sites in S2
 # False Negative (FN) = sites in S1 that are absent in S2 (i.e., "A", "T", "C", or "G" in S1 and "-" in S2)
 # Accuracy = 100 * TP / (TP + FP + FN)
 
 if [[ "${METRIC}" == "accuracy" ]]
 then
     TP=$(cat transposed.txt | awk '$1 == $2 {print}' | wc -l)
-    FP=$(cat transposed.txt | awk '$1 == "-" {print}' | wc -l)
-    FN=$(cat transposed.txt | awk '$1 == "-" {print}' | wc -l)
+    FP=$(cat transposed.txt | awk '$2 != "-" && $1 != $2 {print}' | wc -l)
+    FN=$(cat transposed.txt | awk '$2 == "-" {print}' | wc -l)
 
     echo "Sample,Truth,TP,FP,FN,Result" > "${S2}_result.csv"
     echo -e "${S2}\t${S1}\t${TP}\t${FP}\t${FN}" | awk -v OFS=',' '{print $1,$2,$3,$4,$5, 100 * $3 / ($3+$4+$5)}' >> "${S2}_result.csv"

conf/modules.config

@@ -278,7 +278,7 @@ process {
             enabled: true,
             mode: "${params.publish_dir_mode}",
             path: { "${params.outdir}/validation" },
-            pattern: "*.csv"
+            pattern: "*.{csv,aln}"
         ]
     }
     withName: 'REPORT' {

modules/local/val_gather.nf

@@ -11,6 +11,7 @@ process GATHER {
 
     output:
     tuple val(metric), val(precision_type), path("*.csv"),  emit: result
+    path "*.aln",                                           emit: aln
     path "versions.yml",                                    emit: versions
 
     when: