CRISP-ID: decoding CRISPR mediated indels by Sanger sequencing

Abstract

The advent of next generation gene editing technologies has revolutionized the fields of genome engineering in allowing the generation of gene knockout models and functional gene analysis. However, the screening of resultant clones remains challenging due to the simultaneous presence of different indels. Here, we present CRISP-ID, a web application which uses a unique algorithm for genotyping up to three alleles from a single Sanger sequencing trace, providing a robust and readily accessible platform to directly identify indels and significantly speed up the characterization of clones.

Figure 1. Input, processing and output of the CRISP-ID application.

After uploading a trace file, CRISP-ID draws a chromatogram that displays the sequence peaks and base calls. The user can tweak the base calling by trimming the start and the end of the chromatogram and by adjusting the background cut-off. To obtain the sequences, a “frame primer” is constructed, containing only the homozygous base calls after the start of the spectral shift (A). This frame primer runs over the entire length of the overlapping spectrum. Provided that the frame primer is sufficiently long, it will align with the reference sequence once for each sequence that is present in the mix (B). If no alignments can be found initially, the primer is iteratively trimmed until either alignments are found, or a minimal size of 10 homozygous base calls is reached. Trimming the 3′ end of the frame primer might be necessary due to poor quality base calls near the end of the sequence run, trimming the 5′ end is necessary in case of insertions, and is set to 10 bases by default (not shown in figure). Initially, a “first guess” of the sequences is constructed based on peak height, with the first sequence containing the highest peaks (C). The first sequence is then compared to the reference sequence according to the first frame found in step B. In case of a base mismatch, the base is swapped with the second sequence. During this process of matching the first sequence to the reference sequence, the swapping inherently results in simultaneously solving the second sequence (D). Finally, the user is presented with a multiple alignment of the de-convoluted sequences and the reference sequence, revealing the size and locus of the indels.