Methods for identifying novel gene editing elements
View Patent ↗Embodiments disclosed herein provide methods for identifying new CRISPR loci and effectors, as well as different CRISPR loci combinations found in various organisms. Class-II CRISPR systems contain single-gene effectors that have been engineered for transformative biological discovery and biomedical applications. Discovery of additional single-gene or multi-component CRISPR effectors may enhance existing CRISPR applications, such as precision genome engineering. Comprehensive characterization of CRISPR-loci may identify novel functional roles of CRISPR loci enabling new tools for biomedicine and biological discovery. CRISPR loci have enormous feature complexity, but classification of CRISPR loci has been focused on a small fraction of highly abundant features. Increased genome sequencing has enhanced the sampling of this feature complexity.
1. A method to identify novel CRISPR effector elements, comprising:
training, by a processor, an unsupervised neural network using a set of known CRISPR locus elements that include spacer and repeat elements of the known CRISPR locus, wherein the trained unsupervised neural network comprises hierarchical clustering;
generating, by the processor, and using the training unsupervised neural network, a preliminary set of CRISPR locus classes by separating a set of known CRISPR loci based, at least in part, on a sequence similarity and/or domain similarity between one or more protein elements of the CRISPR loci;
generating, by the processor, a distance matrix data structure based, at least in part, on cumulative similarities between constituent proteins in each CRISPR locus class, wherein the putative CRISPR loci are classified by applying the hierarchical clustering to the distance matrix;
classifying, by the processor, a CRISPR locus using the unsupervised neural network applied to all or a subset of the known CRISPR locus elements as an initial set of inputs; and
identifying, by the processor, putative novel effector elements in the CRISPR locus, and optionally screening each identified putative novel effector element for one or more biological functions.
2. The method of claim 1 , wherein the set of known CRISPR loci are separated based on a domain similarity.
3. The method of claim 2 , wherein the domain similarity is determined using a hidden markov model.
4. A computer program product, comprising:
a non-transitory computer-executable storage device having computer readable program instructions embodied thereon that when executed by a computer cause the computer to identify novel CRISPR effector elements, the computer-executable program instruction comprising:
computer-executable program instructions to train an unsupervised neural network using a set of known CRISPR locus elements that include spacer and repeat elements of the known CRISPR locus elements, wherein the unsupervised neural network comprises hierarchical clustering;
computer-executable program instructions to generate a preliminary set of CRISPR locus classes by separating a set of known CRISPR loci based, at least in part, on a sequence similarity and/or domain similarity between one or more protein elements of the CRISPR loci;
computer-executable program instructions to generate a distance matrix data structure based, at least in part, on cumulative similarities between constituent proteins in each CRISPR locus class, wherein the putative CRISPR loci are classified by applying the hierarchical clustering to the distance matrix;
computer-executable program instructions to classify a CRISPR locus using the unsupervised neural network applied to all or a subset of the known CRISPR locus elements as an initial set of inputs; and
computer-executable program instructions to identify putative novel effector elements in the CRISPR locus, and optionally screening each identified putative novel effector element for one or more biological functions.
5. The computer program product according to claim 4 , wherein the set of known CRISPR loci are separated based on a domain similarity.
6. The computer program product according to claim 5 , wherein the domain similarity is determined using a hidden markov model.
7. A system to identify novel CRISPR effector elements, the system comprising:
a storage device; and
a processor communicatively coupled to the storage device, wherein the processor executes application code instructions that are stored in the storage device and that cause the system to;
train an unsupervised neural network using a set of known CRISPR locus elements that include spacer and repeat elements of the known CRISPR locus elements, wherein the unsupervised neural network comprises hierarchical clustering;
generate a preliminary set of CRISPR locus classes by separating a set of known CRISPR loci based, at least in part, on a sequence similarity and/or domain similarity between one or more protein elements of the CRISPR loci;
generate a distance matrix data structure based, at least in part, on cumulative similarities between constituent proteins in each CRISPR locus class, wherein the putative CRISPR loci are classified by applying the hierarchical clustering to the distance matrix;
classify a CRISPR locus using the unsupervised neural network applied to all or a subset of the known CRISPR locus elements as an initial set of inputs; and
identify putative novel effector elements in the CRISPR locus, and optionally screening each identified putative novel effector element for one or more biological functions.
8. The system according to claim 7 , wherein the set of known CRISPR loci are separated based on a domain similarity.
9. The system according to claim 8 , wherein the domain similarity is determined using a hidden markov model.
10. The system according to claim 7 , wherein the cumulative similarities between constitute proteins are determined, at least in part, by calculating a Euclidean distance between each CRISPR locus.