You can try align_epi_anat.py with the various -partial_coverage options. If datasets are very close to start off, then you can use "-cmass nocmass" instead. If datasets are very far apart, try -giant_move. Output grids can be set for "-master_epi BASE" to give resolution and FOV of the full coverage T1 in this case. The cost function can be changed from the default of lpc to lpa as in the following example.
align_epi_anat.py -anat T1+orig -epi T1_2+orig -base_epi 0 -partial_coverage -master_epi BASE -prep_off -cost lpa -anat_has_skull no -epi_strip None -suffix _al2T1 -AddEdge
If datasets have little overlap at all, then @align_centers may be a useful preprocessing step. Viewing the final results can be done with the @AddEdge script, overlay/underlay features, edge enhancement and checkerboard features in the AFNI GUI.
