ACCURACY OF AUTOMATED DNA-SEQUENCING - A MULTILABORATORY COMPARISON OF SEQUENCING RESULTS

A double-stranded (ds)DNA template of ''unknown'' sequence was distrib uted to approximately 80 core DNA sequencing laboratories by the Assoc iation of Biomolecular Resource Facilities (ABRF) for automated DNA se quence analysis. Forty-fear different facilities responded with 83 usa ble sequence submissions. These sequences were grouped by both sequenc ing protocol (dye-primer or dye-terminator) and whether manually edite d or not. The sequences were aligned with the known sequence, and the number of correct base calls, insertions, deletions, no-calls and misc alls were determined for each group. The dye-primer sequencing protoco l provided the longest and most accurate sequence. The edited dye-prim er data were > 95% accurate out to 400-450 bp, while the edited dye-te rminator data could call only 300-350 bases at this accuracy However 7 5% of the laboratories in this sampling preferred the dye-terminator p rotocol, presumably because of its versatility and convenience. Labora tories that manually edited the automatically called data were able to obtain an additional 100 bases of good sequence when the dye-primer p rotocol was used Surprisingly though, editing of dye-terminator result s did not increase the amount of good sequence, although the dye-termi nator protocol had a superior base-calling ability within the first 10 0 bases of called sequence.