PriSeT: Efficient De Novo Primer Discovery

Marie Hoffmann; Michael T. Monaghan; Knut Reinert

Given the sampled library, we first computed the ground truth by searching directly for the known primer sequences. We used a simple linear text search and marked all data sets having at least one accession with forward and reverse . CC-BY-NC-ND 4.0 International license author/funder. It is made available under a The copyright holder for this preprint (which was not peer-reviewed) is the . https://doi.org/10.1101/2020.04.06.027961 doi: bioRxiv preprint Table 5 . Selected primer sequences for the verification experiment. Sequences are noted in 5' to 3' direction. Melting temperatures are computed using a modified nearest-neighbor method described by Breslauer et al. (1986) assuming a primer concentration of 0.5 µm and salt 50 mM. For ambiguous encodings, the average of both extrema is computed. Sequence parameters violating primer design recommendations are written bold. Critical structures are marked: self-annealing (α), mononucleotid runs (β), CG clamps (γ), exceeding ∆Tm (δ), (A|T)3 tails ( ), and exceeding CG content ranges (γ). (1) or absence (0) bit in the denominator of each cell in Table 7 . We then ran PriSeT with the relaxed constraints listed in the Verification column of Table 6 and searched its result set for the published primers listed in Table 5 . We indicated their presence or absence in the nominator. We could recover all those primer sequences that satisfied the relaxed constraints. The forward primer of DIV4, and both primers of CV did not pass the relaxed Tm filter (above 65 • C) and failed therefore for C s , and the primer pairs E15 and EA failed for C p because of their excessive differences for Tm (more than 8 Kelvin).