The availability of large EST (Expressed Sequence Tag) databases has led to a revolution in the way new genes are cloned. Difficulties arise, however, due to high error rates and redundancy of raw EST data. For these reasons, one of the first tasks performed by a scientist investigating any EST of interest is to gather contiguous ESTs and assemble them into a larger virtual cDNA. The REX (Recursive EST eXtender) algorithm described in this paper completely automates this process by finding ESTs that can be clustered on the basis of overlapping bases, and then assembling the contigs into a consensus sequence. By combining the clustering and assembly steps, REX can quickly generate assemblies from EST databases that are frequently updated without having to preprocess the data. A consensus assembly method is used to correct miscalled bases and remove indel errors. A unique feature of this method is that it addresses the issues of splice variants and unspliced cDNA data. Since REX is a fast greedy algorithm, it can address the problem of generating a database of assembled sequences from very large collections of EST data. A procedure is described for creating and maintaining an Assembled Consensus EST database (ACE) that is useful for characterizing the large body of data that exists in EST databases.