The TM region has been defined in an approximate way as all residues
between the first and the last predicted TMH residue (TMHMM2).
The TM domain sequences have been processed by our automated modeling
pipeline ModPipe, and the results have been loaded into ModBase. The
newest template PDB set has been used to identify possible template
To produce the attached images, I counted residues for each sequences
that have been modeled above a given sequence identity cutoff (all
cutoffs from 20 to 100% sequence identity), and thus calculated the
percentage of model coverage for each sequence. This method takes into
account overlapping models for individual sequences.
The first image (humantmmodels_all.png) show the full dataset, the
second one a part of it, so that details are more visible.
For example (humanthmodels_50.png):
30% sequence identity (darkblue):
310 sequences have been modeled at at least 30% sequence identity with
100% length coverage.
1113 sequence have been modeled at at least 30% sequence identity with
at least 50% length coverage.
100% sequence identity (lightgreen):
17 sequence have been modeled with 100% sequence identity/ 100% length coverage
42 sequence have been modeled with 100%sequence identity/50% length coverage