Multiplex primer prediction software for divergent targets

Gardner, Shea N.; Hiddessen, Amy L.; Williams, Peter L.; Hara, Christine; Wagner, Mark C.; Colston, Bill W.

Here, we outline a greedy algorithm used to calculate conserved sets of multiplexed primers to amplify fragments from each member of a target set of sequences, and provide a more detailed description in the Supplementary Methods. First, we enumerate all candidate oligos fitting user-specified requirements for length, T m , and lack of hairpin formation. We rank pairs of these by the number of targets in which that pair occurs within a distance range d 1 bases to d 2 bases of one another, where these might be specified so as to bound a reasonable range for a downstream characterization method such electrophoretic discrimination of bands, probe hybridization or sequencing. The most frequent pair is selected as primers. The process is repeated for the remaining targets that would not have an amplicon from the first pair, with the added consideration that new primers selected be predicted not to form dimers with other primers already selected, as well as can be predicted based on nearest neighbor thermodynamic predictions (27) , although free energy calculations cannot predict with certainty that primer dimers will be excluded in practice. There is an option to bin primers into reaction subsets, if desired. With binning, primers are added to a bin until that bin contains b primers, at which point a new bin is begun, following the same process. Binning primers in smaller groups avoids exclusion of the most highly conserved oligos because of primer dimer free energy constraints. The universal set of primers is the set of selected primers to amplify all genomes in the target set. The primers within a bin should be multiplexed into a single PCR reaction, but each bin should be run separately. This is a simple binning strategy, and alternative strategies could be employed such as starting a new bin with any primer pair that dimerizes with other previously selected primers regardless of the number of primers in the bin, but this could have the possible disadvantage of bin explosion to numerous singleplex or small multiplex reactions. The graph-based algorithm of MuPlex (28, 29) is another binning strategy that could be incorporated. Output on the number of primers rejected due to the various filters (T m , hairpin free energy, etc.) is printed to standard output after each round of primer pairs are selected and also cumulatively, so a user can monitor if/which parameters might be too stringent. If an alternative set of primers is desired that does not overlap with the set selected, one can replace with N's the subsequences matching the selected primers and their reverse complements in the target sequences and rerun the software with the modified input sequences.