I would argue that PGA is better represented by (Ss/2.5) based on what I've researched than the value indicated in the NEHRP 2009 provisions; which is (SDS/2.5).
It seems to me that the NEHRP incorrectly identifies the PGA (amplified for soil effects) for liquefaction as (SDS/2.5) because the SDS value includes the 2/3 seismic margin appropriate for structural design only.
Still unclear on appropriate value, but I've heard that Section 11.8.3 of ASCE/SEI 7-05 correctly identifies the PGA for liquefaction evaluation as (Ss/2.5), which would then be multiplied by the published site coeficient (Fa) to account for soil amplification effects.