SWISS-PROT data bank

Figure 2. AP X) profiles for proteins for which the 3D structures are known. (A) Photosynthetic reaction center M subunit (rcem rhovi) and (B) bacteriorhodopsin (bacr halha). The AP(X) profile is computed as the preference profiles difference P(A ) , embrane / ( Ocxiramembrane The preference profile for the transmembrane state is smoothed in a window of width 7, while the preference profile for the extramembrane state is not smoothed. Each observed (extracted from the swiss-PROT data bank) and predicted (based on digitalized form of the AP X) profile—see Section 5.3.6) transmembrane segment is denoted by a horizontal line.

Predictions for Proteins for Which, According to the SWISS-PROT Data Bank, Structure and Topology Are Not Known... 

Figure 7. The number of transmembrane (TM) segments of given width in the testing and training data set (Tables 1 and 2) (A) observed (extracted from the swiss prot data bank) (B) predicted with pref 2.0,...

It has been pointed out that the swiss-prot data bank includes a certain number of membrane proteins for which the structure is not quite reliably determined, in part because older methods were used to determine the location and number of transmembrane segments. These older methods are less reliable because they were trained on a smaller sets of proteins. Hence, we suggest that it is necessary to use new and better methods for detecting these defective topologies in order to carry out additional research on them with the aim to obtain their more accurate structures and topologies. In the present report we have shown one of possible ways to carry out such a procedure and this has been illustrated using several examples. In this we also used the positive inside rule, predictions on similar proteins, and predictions obtained by other methods. 

SWISS-PROT data banks. The proteins under investigation are listed in Table 1, together with their major characteristics required for the following calculations. The proteins selected span a wide variety of proteins and features concerning number of chains and AA residues (//chain = 1-180, Naa = 124-29808), molar mass (Af 14-3500 kg mol ), and quite different shapes. In several cases, the values for Naa and M obtained from the PDB and SWISS-PROT files, respectively, differ significantly. Apart from contributions of any ligands, this is caused by the fact that in crystallographic work very often a certain number of AA residues are not resolved. 

A Bairoch, R Apweiler. The SWISS-PROT protein sequence data bank and its supplement TrEMBL m 1999. Nucleic Acids Res 27 49-54, 1999. 

Another major source are the amino acid sequences direcdy derived from protein sequencing. Thousands of such sequences have been detected by the SWISS-PROT curators in publications (or have been directly submitted by researchers to SWISS-PROT) and entered into the database. Protein sequences detected by the NCBI journal scan have also been included. For some proteins the Brookhaven Protein Data Bank (PDB) (Abola et al., 1996) is the only source for the sequence information. The PDB entries are checked regularly, and new SWISS-PROT entries were created whenever necessary. 

Sequence databanks Sequence databanks hold the sequence information of DNA, RNA, protein translations, or verified protein sequences. These data-banks range from simple data depositories such as the GenBank to well-curated, -checked and -annotated databanks such as SWISS-PROT. Thanks to web technology, these exponentially growing databanks can still be managed. 

A. Bairoch and B. Boeckmann, SWISS-PROT protein sequence data bank Current status. Nucleic Acids Res. 22, 3578-3580 (1994). 

---