High-throughput interrogation of programmed ribosomal frameshifting in human cells

Martin Mikl; Yitzhak Pilpel; Eran Segal

To unambiguously identify the variant of origin, a unique 12-mer barcode sequence was placed at the 841 5' end of each variable region. DNA was sequenced on a NextSeq-500 sequencer. We determined 842 for each read (PE150) its bin barcode and its variant barcode and discarded all the reads that could 843 not be assigned to a bin and a library variant of origin. In a second step, for each variant the mapped 844 reads (paired-end, 150 bases for each mate) were aligned to the designed sequence (in the context 845 of the reporter) and only those ones were counted that mapped without a mismatch over the whole 846 length (disregarding the last 5 bases). Reads showing a mismatch, in particular a single nucleotide 847 deletion or insertion, were quantified separately and if they passed the same threshold applied to 848 correct library variants a PRF readout (% GFP fluorescence) was calculated for them. These 849 We applied a number of filters to the raw sequencing data to reduce experimental noise. First, variants 859 with less than 20 reads (perfect matches along the whole length) mapped across bins were removed. 860