The molecular cloning of certain antigens to thyroid autoantibodies present in patients with autoimmune thyroid disease would be of great value in further understanding the pathogenesis of these diseases, and in devising new approaches for their treatment. The ideal method for molecular cloning is to first obtain a partial amino acid sequence of a purified protein and to construct an oligonucleotide probe for screening an appropriate cDNA library. Unfortunately in the case if thyroid autoimmunity, important antigens such as the TSH receptor and the microsomal antigen have not been purified. An alternate approach devised by Young & Davis (1983) is to construct a cDNA library in a vector that expresses the encoded proteins. The library can then be screened with an antibody as a probe. We have constructed cDNA libraries in gt11 using mRNA purified from human, pig and rat thyroid cells. Our experiences in constructing and screening these libraries will be described. The advantages of this system are 1) the protein does not have to be purified, 2) previously unknown antigens may be identified. The disadvantages are 1) lack of specificity with antibody selection, 2) because the cDNA is inserted in the beta-galactosidase gene in the vector the antigen is expressed as a fusion protein. This may disturb the tertiary structure of the antigen and alter its antigenicity, 3) cDNA inserts frequently only contain part of the antigen molecule, and may therefore lack important epitopes; polyclonal antibody may therefore be preferable to monoclonal, 4) only 1 in 6 cDNA inserts will be in the correct reading frame for antigen expression, 5) the expressed protein is not glycosylated.(ABSTRACT TRUNCATED AT 250 WORDS)