Skip to main content
Figure 5 | BMC Bioinformatics

Figure 5

From: Transmembrane helix prediction using amino acid property features and latent semantic analysis

Figure 5

HMM topology used for transmembrane prediction. The architecture models cytoplasmic and extracellular regions in a single state S1 by 4 different feature clusters as a Gaussian mixture model. The interior membrane region, that is the TM segment excluding 5 residue positions each at both ends, is modeled by a single state S12 with 4 feature clusters as a Gaussian mixture model. The transition from non-TM to TM segment is modeled with 5 sequential states on the non-TM side and 5 sequential states on the TM side. States S18-S22 are connected to states S6-S2 respectively as shown, to accommodate short loops between two TM segments. States S11 is connected to S13 to allow accommodation of short TM helices. All the transition states, S2-S11 and S13-S22 are modeled with a single Gaussian feature cluster.

Back to article page