Does phase matter in diploid sequence data?

[LPlough]

I have a data set with diploid seqeunces that were converted to IUPAC ambiguity codes for heterozygotes, e.g. 

Mex1 ATWGGCYACT    is het for A/T at pos 3 and C/T at pos 7

Mex2 ATWGGCYACT     "                                "                       "

Mex3 ATAGGCTACT     is homozygous at pos 3 and 7

When I convert back to data format appropriate for IMa2 , two lines per ind.:

Mex1 ATAGGCTACT

Mex1 ATTGGCCACT   

Mex2 ATAGGCCACT

Mex2 ATTGGCTACT

Mex3 ATAGGCTACT

Mex3 ATAGGCTACT

I lose phase information ( for Mex1, does A in pos 3 go with T in pos 7 or C?). Will this be a problem when running IMA2 or is each variable site treated more or less independently within a sequence? 

Thanks,

Louis