diff --git a/regression.analyses.r b/regression.analyses.r
index 925c77d..99dde49 100755
--- a/regression.analyses.r
+++ b/regression.analyses.r
@@ -401,9 +401,9 @@ res.failed <- subset(res.failed, subset = grepl(Factor, pattern = "(^#.*|^Too (m
 
 dep.var.cats <- sub(dep.vars, pattern = "^.*:", replacement = "")
 dep.var.conts <- sub(grep(dep.vars, pattern = ".*:.*", value = T), pattern = ":.*$", replacement = "")
-# df <- df[df %>% dplyr::select(all_of(c(dep.var.conts, grep(dep.vars, pattern = ":", value = T, invert = T)))) %>% is.na %>% rowSums() == 0, , drop = F]
 
 if(all(sub.vars != "ALL_SAMPLES")) {
+  df <- df[df %>% dplyr::select(all_of(sub.vars)) %>% is.na %>% rowSums == 0, , drop = F]
   df <- df %>% mutate(across(.cols = c(dep.var.cats, sub.vars), .fns = as.factor))
   N <- with(df, aggregate(df %>%
                             dplyr::select(dep.var.cats, dep.var.conts, other.factors), 
@@ -422,12 +422,12 @@ if(all(sub.vars != "ALL_SAMPLES")) {
                                    by = sapply(sub.vars, FUN = function(x) {get(x)}, simplify = F), FUN = function(y) {if(is.factor(y)) {table(y, useNA = "always")} else {summary_f(y)}})) %>%
     as.matrix() %>% as.data.frame(stringsAsFactors = F) %>% mutate(across(.cols = !sub.vars, .fns = as.numeric))
 
-  valid.cols <- foreach(i = colnames(df.summary), .combine = c) %do% {
-     if(i %in% sub.vars) {i} else
-     if(is.na(sum(df.summary[i]))) {i} else
-     if(str_contains(x = i, pattern = sub(dep.var.conts, pattern = "(^.*$)", replacement = "\\1\\."), logic = "OR")) {i} else
-     if(sum(df.summary[i]) >0) {i}
-    }
+valid.cols <- foreach(i = colnames(df.summary), .combine = c) %do% {
+         if(!is.numeric(df.summary[,i])) {i} else
+         if(!grepl(i, pattern = "\\.NA('s)?$")) {i} else
+         if(sum(df.summary[,i]) >0) {i}
+     }
+
   df.summary <- df.summary[valid.cols]
   
   df.summary.fin <- cbind(df.summary, N)