Selecting rx expressions

[uzibalqa]

I am using `rx' for constructing the equivalent of string regexp. Specifically 
I am matching character sequences
occurring at the beginning of words and at the end of words.

Have come up with two versions for each. What versions would be preferable, if 
at all ?

For matching start of words

Version 1

(seq bow
(group
(or (seq "co"
(any "glmnr"))
"coun" "cum")))

Verion 2

(seq word-start
(or (seq "co" (any "glmnr")) "coun" "cum")

For matching end of words

Version A

(seq (group
(or "ley" "ily" "ly")) eow)

Version B

(seq (or "ley" "ily" "ly") word-end )