Back to the HELM page

HELM - Examples and tests


Example 1: Analysis of a simple protein
The 3D structure of the crambin (PDB and NRL3D code: 1crn) reveals the presence of two helical segments, i.e. 7-19 and 23-30. HELM results evidence the presence of helix motifs in these regions.

Example 2: Analysis of an all-beta protein
Barwin is a wound-induced protein from barley, classified as a all-beta structure (very short helices can be recognized but they do not represent a real structural feature). The sequence contains different helix motifs, that evidently do not correspond to helix segments.

Example 3: Analysis of a membrane protein
The vasoactive intestinal polypeptide receptor 1 (SWISSPROT code: P32241) is a membrane protein with seven transmembrane helices. The helix segments reported in the SwissProt entry correspond to different helix motifs found by HELM. It must be considered that no membrane protein was present in the data set used by Aurora and Rose to extract their helix motif classification, therefore, these motifs can be considered not specific for membrane proteins.


The ability of HELM to find correctly the helix motifs has been verified by visual inspection of sequences. Moreover, the accuracy of HELM has been tested by analyzing a set of 1106 protein sequences derived from the pdb_select data set (May 1999 release) and comparing results to the analysis published by Aurora and Rose (1998). These Authors analyzed a similar data set consisting of 274 protein sequences (pdb_select, March 1994 release) and reported the number of occurrence for each helix motif within the protein data set. Differences between these two data sets consist mainly in the number of protein sequences but both should represent a unique set of proteins, without significant sequence identity. A higher occurrence of helix motifs can be expected in the HELM test, because a higher number of protein has been analyzed. Therefore, the occurrence of each motif has been compared between the two data set by calculating the correlation coefficient, which resulted 0.837 and demonstrates a good agreement between the HELM test and the previous analysis by Aurora and Rose.

Click here to see the results of the HELM analysis. For each protein contained in the pdb_select data set (May 1999 release) it is reported the occurrence of each motif.