The datasets for training and testing PureseqTM can be found here. In addition, the Human proteome data is provided. For each dataset, we also generated the prediction results from Phobius, Philius, Topcons2, and PureseqTM.
| Folder name | Description |
|---|---|
| pdbtm_database | ground-truth labeling from PDBTM database. |
| Train_and_Test_result | training and testing set, also include the prediction results from the four methods. |
| TMP_and_nonTMP_result | dataset for discrinating Transmembrane Proteins (TMPs) and non-TMPs. |
| Human_proteome_result | human proteome dataset from UniProt. |
| source_code,util | source code for evaluation and label generation. |
Given a UniProt ID (e.g., Q9UMS5), users may find the prediction result of PureseqTM by the following link:
http://pureseqtm.predmp.com/view.html?id=Q9UMS5_PureTM&name=Q9UMS5
