**v.0.3.**

* `fst_WC84`: bug fix, the function was not properly configured for multi-allelic markers (e.g. microsatellite, and haplotype format from STACKS). Thanks to Craig McDougall for catching this.

DESCRIPTION

@@ -1,8 +1,8 @@
 Package: assigner
 Type: Package
 Title: Assignment Analysis with GBS/RADseq Data using R
-Version: 0.3.2
-Date: 2016-07-20
+Version: 0.3.3
+Date: 2016-09-22
 Encoding: UTF-8
 Authors@R: c(
   person("Thierry", "Gosselin", email = "[email protected]", role = c("aut", "cre")),

R/fst_WC84.R

@@ -374,7 +374,7 @@ fst_WC84 <- function(data,
   # fst function
   compute_fst <- function(x, ci = ci, iteration.ci = iteration.ci, quantiles.ci = quantiles.ci) {
     # x = data.genotyped # test
-
+    # x = data.genotyped %>% filter(POP_ID %in% c("upj", "dsj"))
     # Markers in common between all populations ********************************
     pop.number <- n_distinct(x$POP_ID)
 
@@ -505,19 +505,31 @@ fst_WC84 <- function(data,
     )
 
     data.genotyped.het <- x %>%
-      mutate(het = ifelse(stri_sub(GT, 1, 3) != stri_sub(GT, 4, 6), 1, 0))
-
+      mutate(
+        het = ifelse(stri_sub(GT, 1, 3) != stri_sub(GT, 4, 6), 1, 0),
+        AL1 = stri_sub(GT, 1, 3),
+        AL2 = stri_sub(GT, 4, 6)
+      ) %>% 
+      select(-GT) %>%
+      tidyr::gather(data = ., key = ALLELES_GROUP, value = ALLELES, -c(INDIVIDUALS, MARKERS, POP_ID, het)) %>%
+      select(-ALLELES_GROUP) %>% 
+      group_by(MARKERS, POP_ID, ALLELES) %>%
+      summarise(n = length(het[het == 1])) %>% 
+      group_by(MARKERS, ALLELES) %>% 
+      tidyr::spread(data = ., key = POP_ID, value = n, fill = 0) %>% 
+      tidyr::gather(data = ., key = POP_ID, value = mho, -c(MARKERS, ALLELES))
 
     fst.stats.prep <- suppressWarnings(
       data.genotyped.het %>%
-        group_by(MARKERS, POP_ID) %>%
+        # group_by(MARKERS, POP_ID) %>%
         # = the number of heterozygote individuals per pop and markers
-        summarise(mho = sum(het, na.rm = TRUE)) %>%  
-        group_by(MARKERS) %>%
-        tidyr::spread(data = ., key = POP_ID, value = mho) %>%
-        tidyr::gather(key = POP_ID, value = mho, -MARKERS) %>%
-        mutate(mho = as.numeric(stri_replace_na(str = mho, replacement = 0))) %>%
-        full_join(freq.al.locus.pop, by = c("POP_ID", "MARKERS")) %>%
+        # summarise(mho = sum(het, na.rm = TRUE)) %>%  
+        # group_by(MARKERS) %>%
+        # tidyr::spread(data = ., key = POP_ID, value = mho) %>%
+        # tidyr::gather(key = POP_ID, value = mho, -MARKERS) %>%
+        # mutate(mho = as.numeric(stri_replace_na(str = mho, replacement = 0))) %>%
+        # full_join(freq.al.locus.pop, by = c("POP_ID", "MARKERS")) %>%
+        full_join(freq.al.locus.pop, by = c("POP_ID", "MARKERS", "ALLELES")) %>%
         mutate(
           mhom = round(((2 * nal * P - mho)/2), 0),
           dum = nal * (P - 2 * P^2) + mhom

README.md

@@ -113,6 +113,11 @@ The Amazon image can be imported into Google Cloud Compute Engine to start a new
 ## New features
 Change log, version, new features and bug history now lives in the [NEWS.md file] (https://github.com/thierrygosselin/assigner/blob/master/NEWS.md)
 
+
+**v.0.3.3**
+* `fst_WC84`: bug fix, the function was not properly configured for multi-allelic markers (e.g. microsatellite, and haplotype format from STACKS). Thanks to Craig McDougall for catching this.
+
+
 **v.0.3.2**
 * `assignment_mixture`: added a check that throws an error when pop.levels != the pop.id in strata