The amplified libraries were purified, quantified and pooled in equimolar amounts. The pool was loaded at 4 pM on one MiSeq flowcell. The sequencing was performed sequentially from both ends each for 154 cycles of chemistry and 151 cycles of imaging, to prevent imaging from the tri-nucleotide PCR-specificity adapters. An additional six cycles were used to read the index. The resulting reads were deconvoluted based on their index. The results were analyzed through the identical analysis pipeline, trimming the reads to 122 nucleotides for comparable statistical analysis. The files corresponding to the raw reads are publicly available at the NCBI Short Read Archive (SRP009487.1) [28].
