PMC:99051 / 21698-23356 JSONTXT

{"project":"2_test","denotations":[{"id":"11882253-8601731-10250606","span":{"begin":1539,"end":1541},"obj":"8601731"},{"id":"11882253-6199242-10250607","span":{"begin":1542,"end":1544},"obj":"6199242"}],"text":"In order to explore the limitation of the method we also computed all possible restricted fingerprints of MCP-2 and Cystatin C containing only 5 different restriction enzymes. Here we obtained the same results when we used the five shortest fragments of the fingerprints, but misleading results, if we left out for example the restriction enzyme that produced the shortest fragment. This observation can be easily explained, as partial cDNA sequences derived from oligo dT synthesis can lack the most 5' restriction site and include the most 3' restriction site, but never vice versa. If the numbers of fragments were reduced to the 4 shortest ones, the results were perturbed by unspecific fingerprints that were not derived from the specific cDNAs but were erratically composed of fragments derived from different cDNAs within the pool of 96 colonies. More of those comparisons started to reveal significance limits such as the number of fragments that should be contained in a fingerprint. From the limited amount of comparisons it would seem that having 6 different fragments in a pattern and finding all of them works reliably even with low abundant genes like MCP-2 and Cystatin C that were induced to medium abundance by the treatment (3 to 4 transcripts within the 72.000 analyzed transcripts or about 0.05%).. Finding only 3 to 4 fragments is sufficient only for high abundant genes with a significant differential expression. E.g., serum albumin and SPR1a with abundances of about 0.5% in liver and skin libraries, respectively [26,27], could be identified and quantified by using only 3 restriction enzymes in the corresponding cDNA libraries [28]"}