Methods, models, systems, and apparatus for identifying target sequences for Cas enzymes or CRISPR-Cas systems for target sequences and conveying results thereof
Disclosed are thermodynamic and multiplication methods concerning CRISPR-Cas systems, and apparatus therefor.
1. A method for selecting and producing an engineered CRISPR complex for targeting and/or cleavage of a candidate target nucleic acid sequence within a eukaryotic cell, comprising the steps of:
(a) determining amount, location and nature of mismatch(es) of guide sequence of potential CRISPR complex(es) and the candidate target nucleic acid sequence,
(b) determining contribution of each of the amount, location and nature of mismatch(es) to hybridization free energy of binding between the target nucleic acid sequence and the guide sequence of potential CRISPR complex(es) from a training data set,
(c) based on the contribution analysis of step (b), predicting cleavage at the location(s) of the mismatch(es) of the target nucleic acid sequence by the potential CRISPR complex(es),
(d) selecting the CRISPR complex from potential CRISPR complex(es) based on whether the prediction of step (c) indicates that it is more likely than not that cleavage will occur at location(s) of mismatch(es) by the CRISPR complex;
(e) producing the selected CRISPR complex or nucleic acid molecule(s) encoding the selected CRISPR complex for targeting and/or cleavage of the candidate target nucleic acid sequence within the eukaryotic cell; and
(f) delivering the selected CRISPR complex or nucleic acid molecule(s) encoding the selected CRISPR complex into the eukaryotic cell, wherein the selected CRISPR complex targets and/or cleaves the candidate target nucleic acid sequence within the eukaryotic cell.
2. The method of claim 1 wherein the candidate target sequence is a DNA sequence, and the mismatch(es) are of RNA of potential CRISPR complex(es) and the DNA.
3. The method of claim 1 , wherein step (b) is performed by
determining known local free energies, ΔGij(k), between every guide RNA sequence i and target DNA nucleic acid sequence j at position k,
calculating values of the effective free-energy Z ij using the relationship p ij ∝e −βZij ,where p ij is measured cutting frequency by guide RNA sequence i on target DNA nucleic acid sequence j in the training set and β is a positive constant of proportionality,
determining the weights which are position-dependent weights α k by fitting the known value of ΔGij(k) and the calculated value of Z ij across each guide RNA/target DNA sequence pair in the training set in the sum across all N bases of the guide-sequence
Z
ij
=
∑
k
=
1
N
α
k
Δ
G
ij
(
k
)
by writing the above equation in the matrix form {right arrow over (Z)}=G{right arrow over (α)}
and wherein, step (c) is performed by
estimating the effective free-energy Zest using the determined position dependent weights in the equation
{right arrow over (Z est )}=G {right arrow over (α)}
and determining estimated spacer-target cutting frequencies p est αe −βZest , to thereby predict cleavage.
4. The method of claim 2 wherein the distance, in bp, between the first and last base of the target sequence is 18.
5. The method of claim 1 wherein predicting cleavage comprises predicting whether cleavage is more likely than not to occur at location(s) of mismatch(es), and thereby predicting cleavage.
6. The method of claim 1 , further comprising normalizing the calculated values of the effective free energy of hybridization Z for each guide RNA/target DNA sequence pair in the training set.
7. The method of claim 1 , further comprising filtering out calculated value of the effective free energy of hybridization Z for each guide RNA/target DNA sequence pair in the training set which have a sequencing depth which is below a minimum sequencing depth.
8. The method of claim 1 , wherein the method is implemented by a computer system comprising:
a. a memory unit configured to receive and/or store sequence information of the candidate target nucleic acid sequence; and
b. one or more processors alone or in combination programmed to perform steps (a) to (d).
9. The method of claim 1 , wherein step (b) is performed by:
defining a thermodynamic model having a set of weights linking effective free energy of hybridization Z to local free energies G;
defining a training set of the guide sequence/target DNA sequence pairs;
inputting known values of local free energies G for each guide sequence/target DNA sequence pair in the training set;
calculating a value of effective free energy of hybridization Z for each guide sequence/target DNA sequence pair in the training set;
determining the weights using a machine learning algorithm, and
outputting the weights whereby the weights can be used to estimate the free energy of hybridization for any sequence.
10. The method of claim 1 wherein the guide sequence is comprised within a single guide RNA (sgRNA) or within a CRISPR-Cas system chimera RNA (chiRNA).
11. The method of claim 1 , wherein the selected CRISPR complex generates a cleavage within the candidate target nucleic acid sequence within the eukaryotic cell.