In my notes I found two genitive formations for nouns ending in ie - sindieo (PE17/72) and máriéno (PE17/59). The first example implies normal rules apply. The second is a bit more interesting, indicating that a "joining consonant" might be used in some conditions.

As for the Tengwar, I'd guess you'd use yanta for post-vocalic "y", since the difference between y and diphthongal i is pretty small anyway.

+Paul Strack
according to +Måns Björkman
When occurring solely, and not following a consonant, the palatal sound y was represented by the silent tengwa anna together with the palatalization tehta.

on http://at.mansbjorkman.net/teng_quenya.htm

+Paul Strack As I see, máriéno is rather a late stage of an experiment, 'what if laurië is an adverb, then what is márie?' So it can be excluded after RGeO. However, the second pne is disturbing. I now see that before 1936, Tolkien uses tyalio consistently and tyalyo once.

+Александр Запрягаев
why is it disturbing? It's not awkward looking and not  difficult to pronounce.

+Tamas Ferencz Well, it is much more awkward than the proposed tyalio, with three vowels in an epic hiatus. Especially if pronounced as written, with i stressed. Tyaliéo is altogether different thing, that I can relate to. Also, this solution does not generalize to other clusters; I cannot make a form like mátieuva with sense.

I missed anna altogether; much thanks! I've decided to study Outline of Phonology carefully now; how much it hurts that the cluster development should be considered much wrong (tengwa etc.)!

I use máriéno, but that's rather subjective.

+Ицхак Пензев I thought máriéno appeared in a bit which treats laurië as adverb, not adjective plural, which is superseded by RGeO printed canon. The same texts proposes a further nasty deletion blasting away the -ndon adverbials, so I tend to disregard it. I'm for PE21 here strangely enough.