Multiplex primer prediction software for divergent targets

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

We also generated primer sets for several species with high sequence diversity: HIV-1, FMDV, Norwalk virus and influenza A segments HA and NA (summary in Tables 2 and 3, primer sequences available as Supplementary Data). MPP parameters were the same as in Table S1 , with T m = 35-50 C for 10-mer primers and T m = 55-70 C for 17-18-mer primers, and sequences were downloaded in 2007 or 2008 (influenza only). To illustrate the challenges of designing primers from an alignment, we aligned these organisms using MUSCLE (30) when possible. For the HIV-1 and influenza A segments HA and NA, MUSCLE ran out of memory before completing. An alternative alignment tool, ClustalW (31), had completed only a small fraction of the alignment after running for days. Therefore, for these large data sets, a random selection of $35 sequences for each target was aligned with MUSCLE, and this alignment was used to build a profile Hidden Markov Model (HMM) (hmmbuild) using HMMer (http://hmmer.wustl.edu/). (32) The full sequence set was then aligned to the HMM using hmmalign. For Norwalk virus and FMDV, we designed multiplex-degenerate primer sets using GreeneSCPrimer (20) and HYDEN (22) , but the other three targets sets were too large for those programs. None of the other primer prediction programs we tried (as indicated in the 'Results' section) would scale to handle even these two smallest target sets.