Nucleic Acids Research Group Report

The Nucleic Acids Research Group of the ABRF is now completing a study of strategies used for DNA sequencing primer design and of the efficacy of these primers. This study was conducted due to the overall paucity of rigorous empirical studies of sequencing primer design strategies and criteria. Interested ABRF members were provided a 300-base 'test sequence' that was selected for its known sequencing characteristics. Participants were requested to provide their preferred primer design strategies, including specific biophysical characteristics of the primers. In addition, participants were given the opportunity to submit as many as three primers selected from the 'test sequence' using their preferred primer design strategies.

The committee analyzed the biophysical characteristics of each of the submitted primers. Each of the primers was then synthesized, purified, and used in a standard sequencing reaction for the Perkin Elmer/ABI 377 Prism Automated DNA Sequencer. The quality of the sequence ladders generated by each primer was assessed by trimming the data to remove ambiguous base calls on both the 5' and 3' termini, alignment of the experimental data with the known sequence, and by tallying the number of sequence errors.

In parallel with the analysis of the submitted primers, a panel of 18-base control primers spaced evenly across the 300-base 'test sequence' was designed, synthesized, and purified. These 95 primers were chosen with 5' termini designed to hybridize at three (3) base intervals across the test sequence. As for the submitted primers, each of these primers was used in standard sequencing reactions to provide baseline data on the function of randomly selected primers; i.e., primers that are not selected according to any specific criteria.

The results of our poll of primer design strategies showed that most investigators favor oligos of 18-24 bases, 40-60 %G+C, Tm of 53-65°C, a 1-2 base 3' GC clamp, hairpin stems of less than 2-3 bp, homopolymeric runs of less than 4-5 bases, and secondary priming sites of less than 3-4 bp. Physicochemical analysis of the submitted primers showed that, as would be expected, the measured characteristics of the majority very closely matched the reported ideal characteristics. Surprisingly, however, a fraction of the submitted primers fell significantly outside the idealized ranges, suggesting that some investigators may inadvertently or purposefully select 'nonoptimal' primers. Finally, slightly more than half of all submitted primers were selected 'manually', without the aid of any primer selection software programs currently available.

The sequencing data also presented some unexpected results. First, all of the submitted primers, and nearly all of the control primers, yielded very good sequencing read lengths and accuracy; i.e. greater than 700 bases with less than 1-2% error. Moreover, there was no evidence that the sequence quality was affected to a large degree by primer length, %G+C or Tm, primer secondary structure, 3' GC clamps, homopolymeric runs, or secondary priming sites. Although several primers did fail to sequence well, these primers were invariably among the random controls, and exhibited extreme characteristics such as 7-8 base 3' homopolymeric sequences. Even those submitted primers which fell significantly outside idealized norms for the physicochemical characteristics examined provided excellent sequence data.

Our preliminary conclusions from this study suggest that sequence primer performance is unexpectedly insensitive to physicochemical characteristics normally thought to adversely impact sequence results. We are now finalizing this work for submission for publication.

Return to the The ABRF Home Page