Heavy-atom Database System (HATODAS) is a WWW-based tool designed to assist the heavy-atom derivatization of proteins. The conventional procedure for the preparation of derivatives is usually a time-consuming `trial-and-error' process. The present program provides a solution for this problem using a database of known heavy-atom derivatives. A database search suggests potential heavy-atom reagents for any target protein based on its amino-acid sequence and crystallization conditions. A mining of the database identified 93 preferred motifs for heavy-atom binding. The motifs are observed frequently at the actual heavy-atom-binding sites encountered in the process of structure determination.