AbstractSplice variants play an important role within the cell in both increasing the proteome diversity and in cellular function. Splice variants are also associated with disease states and may play a role in their etiology. Information about splice variants has, until now, mostly been derived from the primary transcript or through cellular studies. In this study information from the transcript and other studies is combined with tertiary structure information derived from homology models. Through this method we have determined that it is possible to effectively model splice variants. Forty models of splice variants for fourteen proteins were produced. Analysis of the models shows that deletions produce superior model validation values. Additions to sequences where there is little homology become increasingly difficult to model with increasing sequence length. Many of the splicing events are associated with post‐translational modification either in the N‐terminal region by changing the signal peptide or by affecting the number or availability of glycosylation sites. Often the alternative exon combinations are associated with loss or gain of whole structural units, as opposed to just changing small loop regions. Losing part of the secondary structure may destabilize neighboring parts of the same secondary structure. Detailed analysis is given of four biomedically relevant proteins (Beta‐site Amyloid Precursor Protein Cleaving enzyme (BACE), Interleukin‐4, Frataxin and Hereditary hemochromatosis protein) and their associated splice variant models. The visualization of these possible structures provides new insights about their functionality and the possible etiology of associated diseases. Proteins 2004;54:000–000. © 2003 Wiley‐Liss, Inc.