Transmembrane Topology Prediction (ProtTM)

ProtTM tool predicts protein transmembrane topology and supports both transmembrane alpha-helix and beta-barrel systems, without species restrictions. It accepts multiple FASTA sequences and is suitable for batch pre-screening of candidate proteins, membrane protein expression planning, and joint evaluation of signal peptides with transmembrane regions.

The model output includes global class prediction, residue-level topology labels, transmembrane segment boundaries, and signal peptide cleavage sites. Note: this tool is intended for topology inference and is not a direct subcellular localization predictor.

1. Protein Sequences (up to 10 FASTA entries):

Parsed sequences: 0, total residues: 0



Label Definitions (Residue Level)

  • B: Beta Strand (TM), transmembrane beta-sheet strand
  • I: Inside (Cytoplasm), inner/cytoplasmic side of membrane
  • O: Outside (Extracellular), outer/extracellular side of membrane
  • M: Alpha Helix (TM), transmembrane alpha-helix
  • P: Periplasm, periplasmic space region
  • S: Signal Peptide, signal peptide region

Model Performance Metrics

====================================================================================
BENCHMARK REPORT (Filtered Test Set) [Ensemble: 3 models, method=vote]
====================================================================================
Total Residues Evaluated: 2500818
Overall Accuracy:         0.9855
Overall MCC:              0.9188
Balanced Accuracy:        0.9185
Macro Precision/Recall/F1:0.9299 / 0.9185 / 0.9238
Weighted P/R/F1:          0.9875 / 0.9855 / 0.9865
Micro P/R/F1:             0.9875 / 0.9855 / 0.9865
Classes absent in y_true for this benchmark split: P
Mean Sequence Accuracy:   0.9756
Median Sequence Accuracy: 1.0000
Sequence Exact Match:     0.8535

[Per-Protein Classification]
Protein Accuracy:         0.9879
Protein MCC:              0.9621
Protein Macro P/R/F1:     0.8878 / 0.9375 / 0.9025
Type       | Support  | Precision  | Recall     | F1        
------------------------------------------------------------------
GLOBULAR   | 4044     | 0.9948     | 0.9953     | 0.9951    
SIGNAL     | 273      | 0.9672     | 0.9707     | 0.9689    
TM         | 566      | 0.9715     | 0.9629     | 0.9672    
SP+TM      | 15       | 0.5600     | 0.9333     | 0.7000    
BETA       | 63       | 0.9455     | 0.8254     | 0.8814    

[Signal Peptide Metrics]
SP Presence Acc/MCC:      0.9966 / 0.9715
SP Presence P/R/F1:       0.9565 / 0.9904 / 0.9731
CS Exact Accuracy (true SP):   0.8103
CS Relaxed@+/-2 (true SP): 0.9196
CS Exact P/R/F1:          0.7826 / 0.8103 / 0.7962
CS Relaxed P/R/F1:        0.8882 / 0.9196 / 0.9036

[Per-Segment Metrics]
Segment  | TP     | FP     | FN     | Precision  | Recall     | F1         | BoundaryMAE
----------------------------------------------------------------------------------------------------
B        | 763    | 68     | 48     | 0.9182     | 0.9408     | 0.9294     | 0.9463    
M        | 2544   | 1129   | 483    | 0.6926     | 0.8404     | 0.7594     | 1.7750    
S        | 292    | 38     | 19     | 0.8848     | 0.9389     | 0.9111     | 0.1284    
TM       | 3307   | 1197   | 531    | 0.7342     | 0.8616     | 0.7929     | 1.5838    

[Per-Class Performance (Residue Level)]
Class    | Support  | Precision  | Recall     | F1         | MCC        | Specificity
------------------------------------------------------------------------------------------------
B        | 7210     | 0.8504     | 0.8513     | 0.8508     | 0.8504     | 0.9996    
I        | 2263185  | 0.9953     | 0.9931     | 0.9942     | 0.9395     | 0.9549    
M        | 64393    | 0.9172     | 0.8404     | 0.8771     | 0.8749     | 0.9980    
O        | 158275   | 0.9129     | 0.9433     | 0.9278     | 0.9230     | 0.9939    
P        | 0        | 0.0000     | 0.0000     | 0.0000     | 0.0000     | 0.9980    
S        | 7755     | 0.9737     | 0.9645     | 0.9691     | 0.9690     | 0.9999    

[Confusion Matrix Counts]
           B      I      M      O      P      S
   B    6138    181      0    486    405      0
   I     247 2247605   2794   9067   3370    102
   M       0   5705  54115   4535      0     38
   O     833   4742   2067 149306   1265     62
   P       0      0      0      0      0      0
   S       0     85     24    165      1   7480

[Confusion Matrix Row-Normalized]
           B      I      M      O      P      S
   B   0.851  0.025  0.000  0.067  0.056  0.000
   I   0.000  0.993  0.001  0.004  0.001  0.000
   M   0.000  0.089  0.840  0.070  0.000  0.001
   O   0.005  0.030  0.013  0.943  0.008  0.000
   P   0.000  0.000  0.000  0.000  0.000  0.000
   S   0.000  0.011  0.003  0.021  0.000  0.965
====================================================================================
        

Last updated: 2026-05-25