Sequencing or haplotype data are potentially problematic because they
usually violate the assumption of no LD between sites. However, if
you have data from enough independent regions this may still allow
accurate inference (eg see example from Conrad et al Nature Genetics
2006, Supp Info or Falush et al, Science 2003).
To enter sequence/haplotype data, simply treat each segregating site
as an independent locus and use usual Structure input format.
jonathan