Design strategies and performance of custom DNA sequencing primers.

ABSTRACT

This study surveyed strategies of sequencing primer selection and evaluated primer performance in automated DNA sequencing. We asked participants to relate their preferred primer design strategies to identify primer characteristics that are considered most important in sequencing primer design. The participants preferred primers of 18-24 nucleotides (nt), 39%-58% G + C, a melting temperature (Tm) of 53 degrees-65 degrees C with a 1-2 nt 3' GC clamp, hairpin stems of less than 2-3 bp, homopolymeric runs of less than 4-5 nt, primer dimers of less than 3-4 bp and secondary priming sites of less than 3-4 bp. We provided a 300-bp test sequence and asked participants to submit sequences of 1-3 optimal sequencing primers. Submitted primers ranged from 17-24 nt and largely conformed to the preferred parameters. Submitted primers were distributed across the test sequence, although some sites were disfavored. Surprisingly, approximately 45% of the primers were selected "manually", more than by any software package. Each of 69 submitted and 95 control primers, distributed at 3-bp intervals across the test sequence, were synthesized, purified and tested using a Model 377 PRISM DNA Sequencer with dichlororhodamine dye terminator reagents (dRhodamine dye terminators). Approximately half of the control primers were also tested using rhodamine dye terminator reagents ("old" rhodamine dye terminators). The results indicated that primer physico-chemical characteristics thought to have a strong impact on sequencing performance had surprisingly little effect. Thus, primers with high or low percent G + C or Tm, strong secondary priming scores or long 3' homopolymeric stretches yielded excellent sequences with the dRhodamine dye terminator reagents, although these characteristics had a stronger effect when the old rhodamine reagents were used. The old rhodamine reagents gave sequences with a similar average read length, but the number of errors and ambiguities or "N's" was consistently higher. Moreover, the effects of the primer physico-chemical characteristics were also more evident with the old rhodamine dyes. We conclude that under optimal sequencing conditions with highly pure template and primer, many of the commonly applied primer design parameters are dispensable, particularly when using one of the new generation of sequencing reagents such as the dichlororhodamine dye terminators.