by Isidore Rigoutsos Labtimes 04/2015
Off-target effects are a major concern in gene editing with the CRISPR-Cas system. A new web-based software tool may help to determine off-target sites for guide RNA.
Off-Spotter takes a close look at CRISPR-Cas off-target sites and identifies genomic locations identical to the guide RNA.
The “clustered regularly interspaced palindromic repeats” system (CRISPR) is a two-component immune-response mechanism that many prokaryotes have developed and which implements an adaptive resistance mechanism to infections by viruses and plasmids.
There are two basic elements to the system. The first is a cluster of genomic repeats (CRISPR loci) interrupted by “spacers” (or, “protospacers”), which are short sequence fragments derived from elements foreign to the prokaryotic genome at hand. The second comprises the “CRISPR-associated” (Cas) nuclease genes that are located close to the CRISPR loci. The sequences of the CRISPR loci are transcribed into short RNAs that guide the Cas nucleases to the foreign DNA sites that will be cleaved. Soon after its discovery, this powerful system was adapted into a controlled tool for genome engineering uses.
An experimental implementation of the CRISPR-Cas system begins with the design of a gRNA, a guide RNA sequence that is complementary to the intended DNA target. Only instances of a gRNA that are followed by a “protospacer adjacent motif” or PAM (typically, a sequence such as NGG, NAG, NNNNACA, etc.) can be targeted. Even though it is relatively straightforward to come up with such gRNA sequences, not all gRNAs are equally “well-behaved”: some can potentially target many unintended genomic locations in addition to the desired site. Thus, these so-called “off-targets” are an important consideration during gRNA design.
The ideal gRNA sequence is one that has few (or, ideally zero) identical or near-identical instances on the genome beyond the one that is the intended target. It should be pointed out, however, that the mere existence of a gRNA sequence (+ PAM) in another location of the genome does not necessarily mean that this ‘other’ location will be targeted as well. The rules that determine whether a gRNA+PAM location, intended or unintended, will in fact be targeted are poorly understood and currently the focus of active investigations (Doench et al., Nat. Biotechnol. 32(12):1262-67).
Nonetheless, it is important to know ahead of time for a given candidate gRNA sequence, whether the target genome contains one, few, or many other locations where the gRNA can also be found identically or near-identically.
This is a very practical concern given that, as our group has shown before, in the human, mouse and other genomes there are numerous short DNA sequences, with numerous instances in messenger RNAs, intronic space and intergenic space (Rigoutsos, I., Current Biology, 20(3), 110-13). Even essential molecules, such as tRNAs, may not be immune from such considerations (Telonis, et al., RNA Biology, 12(4), 375-80).
This is where Off-Spotter comes in. Venetia Pliatsika and myself from the Thomas Jefferson University, USA, recently described the Off-Spotter algorithm and made available an implementation (source code) as well as a web-based interface for interactive use by practitioners (Pliatsika & Rigoutsos, Biology Direct, 10(1), 1-10).
At the beginning of June 2015, we also released a utility that permits users to access the system in an unsupervised “batch mode” that can process thousands of input sequences at a time. Off-Spotter is our contribution to the list of tools that are available for determining “off-target” sites for guide RNAs in the CRISPR-Cas nuclease system. The web interface, the source code and the batch utility are accessible from the website of Thomas Jefferson University’s Computational Medicine Center at http://cm.jefferson.edu/Off-Spotter.
For a given candidate gRNA sequence and associated PAM, Off-Spotter guarantees to identify and exhaustively report all genomic locations that are identical to the gRNA or differ from it in fewer than a user-controlled number of positions. In addition, the user is given explicit control with regard to defining in real-time (prior to each run) the location and number of nucleotides that comprise the “seed”. For each of the reported candidate off-target sites, Off-Spotter also reports any known annotation that may be available for that location of the genome together with a hyperlink to the same location on the University of California Santa Cruz genome browser (http://genome.ucsc.edu).
What sets Off-Spotter apart is that it can generate all of this information in a matter of seconds, even for gRNAs that may have tens or hundreds of thousands of potential off-targets. Off-Spotter achieves its speed in software, solely through algorithmic means, and does not rely on hardware accelerators of any kind.
Last Changed: 08.07.2015