how to pick fragments for a transmembrane protein

Dear all,

I have created fragments with make_frags.pl script using the following command lne:
make_fragments.pl -verbose -nosam -nopsipred -nojufo -samfile UapA_SAM.ss2 -id UapA_ UapA.fasta

where UapA_SAM.ss2 is the SAM-predicted secondary structure in pripred_ss2 format, created in the following way:
ss_pred_converter.py --sam UapA.rdb > UapA_SAM.ss2

The UapA.rdb file has been obtained from ROBETTA server. Is this OK?

The default weight file used for fragment picking by "make_fragments.pl" script is the following:
# score name priority wght min_allowed extras
ProfileScoreL1 700 1.0 -
ProfileScoreStructL1 100 1.4 -
SolventAccessibility 500 0.5 -
Phi 300 3.9 -
Psi 200 0.9 -

No SecondarySimilarity and RamaScore are used. Are these priorities and weights suitable for picking fragments for membrane proteins? If not, could someone please recommend me some modifications?

Thanks in advance,
Thomas