Sequence analysis
Sequence files were retrieved from the MiSeq illumina platform as a raw FASTQ file. Sequences corresponding to the coding regions were extracted using a pattern search program written by Perl script. Only the sequences with an 8 nucleotide perfect match upstream of the start codon and downstream of the random region are considered as reliable for downstream analysis. The reliable sequences (in FASTQ format) were further processed using FASTAptamer-count (Alam, Chang, & Burke, 2015) to rank sort sequences based on their read counts. Consensus sequence logos were created using WebLogo 3 (Crooks, Hon, Chandonia, & Brenner, 2004) based on the top 50 reads with specific conditions (fixed letter and positions).