[modeller_usage] Modeling large protein sequence having homologues with less similarity.

Hello,            My target sequence of protein is too large containing more than 1000 amino acids and I found 4 homologues in PDB showing very less similarity i-e. two showing 28% similarity, one showing 24% and last one showing 19% similarity with my target sequence. Also the starting residues of templates have no homology with my target sequence and Modeller has to read from 1st residue of template, that bocomes impossible in this case. So now how can I model my target sequence?