Update susiF.R

Author: william-denault

R/susiF.R

@@ -64,8 +64,10 @@
 #'
 #' @param L_start number of effect initialized at the start of the algorithm
 #'
-#' @param nullweight numeric value for penalizing likelihood at point mass 0 (should be between 0 and 1)
-#' (useful in small sample size)
+#' @param nullweight numeric value for penalizing likelihood at point mass 0. This number roughly corresponds
+#' to the number of zeros observation you add  (useful in small sample size). Default is 10 as recommended by Stephens in
+#' False discovery rate a new deal. Setting it too low tend lead to adding false discoveries. Setting it too
+#' high may reduce power. 
 #'
 #' @param thresh_lowcount numeric, used to check the wavelet coefficients have
 #'  problematic distribution (very low dispersion even after standardization).