problems with input files?

Running I-tasser locally is telling me there are problems with a handful of my input files (about 6 out of 350 have failed to simulate).

All of the protein sequence fastas I used (of the 350) were created in exactly the same way and are valid fasta files as I use them for many other things.

This is a sample of the error I get:

nv-PPAS warning: GLY ne CYS for res=128, convert GLY to CYS
Env-PPAS warning: GLY ne CYS for res=127, convert GLY to CYS
Env-PPAS warning: GLY ne CYS for res=128, convert GLY to CYS
Env-PPAS warning: GLY ne CYS for res=127, convert GLY to CYS
Env-PPAS warning: GLY ne CYS for res=128, convert GLY to CYS
Env-PPAS warning: GLY ne CYS for res=127, convert GLY to CYS
Env-PPAS warning: GLY ne CYS for res=128, convert GLY to CYS
Env-PPAS warning: GLY ne CYS for res=127, convert GLY to CYS
Env-PPAS warning: GLY ne CYS for res=128, convert GLY to CYS
type= easy, n_good=22, M0= 8, M1_for_medm=1.6, n_sg= 8 role=SEQ
domain2=no
type=easy----n_temp_use_for_restraints=20
4.1 run simulation
127 CYS ne GLY !!!!!!
128 CYS ne GLY !!!!!!
init.dat is wrong!!!!! 1 - 4hrzA CYS ne GLY MUSTER
init.dat is wrong!!!!! 1 - 4hrzA CYS ne GLY MUSTER
You have 4 errors in your input files. Please fix these problems before you run TASSER

There are problems with your input files
Please check before simulation

And here is the content of the fasta itself, can you explain why it's failing? Is it a problem with the PDB file 4hrz that the algorithm is picking out?

>PAK_01889 PAK_01889 baseplate wedge subunit 2122669:2123055 reverse MW:14459
MTSQILTNIYGCGWKFPPQFSLETGIAMAEGAENVRQSMKILFLTEPGERIMREDYGCGL
NDYMFENISDELLSEIQTRIEERVLRYEPRAEITDIQITQKTDSPNTLHIQVTYALRGSQ
ISQQLEXX

I'm guessing it's a problem with the undertermined amino acids "XX", but of the 6 proteins tthat are failing, only 4 of them contain X residues, so it seems that might not be the whole story