Systems and methods for synthetic regulatory sequence design or production
The technology described herein is directed to systems and methods for synthetic regulatory sequence design or production. In several aspects described herein are methods of designing and optionally synthesizing shortened or otherwise modified polynucleotide sequences from polynucleotide sequences with transcriptional regulatory functionality. In other aspects described herein are isolated nucleic acid modules and viral vectors, comprising said shortened or otherwise modified polynucleotide as designed or synthesized herein.
1 . A method of synthesizing an adeno-associated virus (AAV) viral vector comprising a synthetic polynucleotide with transcriptional regulatory functionality, the method comprising the steps of:
a) receiving a set of genetic data comprising a regulatory polynucleotide sequence, optionally having a known transcriptional regulatory functionality;
b) processing the set of genetic data into a set of sequence elements that each comprise a portion of the polynucleotide sequence; wherein each of the set of sequence elements is a k-mer fragment of the polynucleotide sequence;
c) processing the set of sequence elements with a machine learning model to determine a transcriptional regulatory score associated with each of the set of sequence elements;
d) identifying at least one sequence element of the set of sequence elements comprising a trough of at least 20 nucleotides between peaks with a transcriptional regulatory score below a threshold as a non-regulatory sequence element;
e) designing a modified polynucleotide molecule based on the determined transcriptional regulatory scores and/or identified non-regulatory sequence elements; wherein at least 10% of the nucleotides within the trough have been removed in the modified polynucleotide molecule; and
f) synthesizing an AAV viral vector comprising the modified polynucleotide molecule.
2 . The method of claim 1 , wherein each of the k-mer fragments comprises a gapped k-mer fragment.
3 . The method of claim 2 , wherein the gapped k-mer fragment comprises a 10 nucleotide-long fragment comprising 4 gaps.
4 . The method of claim 1 , wherein the machine learning model is produced by a support vector machine.
5 . The method of claim 4 , wherein the set of sequence elements are input into the support vector machine.
6 . The method of claim 4 , wherein the support vector machine comprises a gapped k-mer kernel.
7 . The method of claim 4 , wherein the support vector machine comprises a k-DNF k-mer kernel.
8 . The method of claim 1 , wherein the machine learning model is trained using a second set of genetic data comprising a second set of sequence elements that comprise portions of transcriptional regulatory elements and non-regulatory elements.
9 . The method of claim 1 , wherein the modified polynucleotide sequence is designed based on the polynucleotide sequence with the at least one non-regulatory sequence element removed.
10 . The method of claim 1 , wherein the modified polynucleotide sequence is designed based on the polynucleotide sequence with the at least one non-regulatory sequence element replaced.
11 . The method of claim 1 , wherein the modified polynucleotide sequence is designed based on at least one mutation of nucleotide position(s) to alter a transcriptional regulatory score.
12 . The method of claim 1 , wherein synthesizing the shortened modified polynucleotide molecule comprises de novo DNA synthesis.
13 . The method of claim 1 , wherein the known transcriptional regulatory functionality:
comprises transcriptional activation and/or transcriptional repression;
is in any given cell or tissue type;
is in liver cells, cardiac muscle cells, and/or skeletal muscle cells; or
comprises promoter activity.
14 . The method of claim 13 , wherein promoter activity comprises a level of expression of a particular set of genes.
15 . An AAV viral vector comprising a modified polynucleotide molecule that was synthesized using a method comprising the steps of:
a) receiving a set of genetic data comprising a regulatory polynucleotide sequence, optionally having a known transcriptional regulatory functionality;
b) processing the set of genetic data into a set of sequence elements that each comprise a portion of the polynucleotide sequence, wherein each of the set of sequence elements is a k-mer fragment of the polynucleotide sequence;
c) processing the set of sequence elements with a machine learning model to determine a transcriptional regulatory score associated with each of the set of sequence elements;
d) identifying at least one sequence element of the set of sequence elements comprising a trough of at least 20 nucleotides between peaks with a transcriptional regulatory score below a threshold as a non-regulatory sequence element;
e) designing a modified polynucleotide molecule based on the determined transcriptional regulatory scores and/or identified non-regulatory sequence elements; wherein at least 10% of the nucleotides within the trough have been removed in the modified polynucleotide molecule; and
f) synthesizing an AAV viral vector comprising the modified polynucleotide molecule.
16 . A method of synthesizing an AAV viral vector comprising a polynucleotide sequence with transcriptional regulatory functionality that is altered in size without losing transcriptional regulatory functionality comprising the steps of:
a) using a machine to process a set of genetic data comprising polynucleotide sequences to identify putative promoter sequences or putative enhancer sequences;
b) processing the set of putative promoter sequences into a set of sequence elements that each comprise a portion of the putative promoter sequences with a machine learning model to determine a transcriptional regulatory score associated with each of the nucleotides within the putative promoter sequences to identify peaks and troughs based upon the scores provided to each nucleotide within the putative promoter sequences; wherein each of the set of sequence elements is a k-mer fragment of the polynucleotide sequence;
c) identifying at least one member of the set of putative promoter sequences where there is a trough of at least 20 nucleotides between peaks, wherein the scores for members of the trough are below a pre-determined threshold; and
d) synthesizing an AAV viral vector comprising a shortened polynucleotide molecule, wherein at least 10% of the nucleotides within the trough have been removed as compared to the selected identified putative promoter sequence of step (c).
17 . A method of synthesizing an AAV viral vector comprising a synthetic polynucleotide with transcriptional regulatory functionality, the method comprising the steps of:
a) receiving a set of genetic data comprising a regulatory polynucleotide sequence, optionally having a known transcriptional regulatory activity;
b) identifying an enhancer portion of the polynucleotide sequence using a machine learning model based on k-mers;
c) processing the enhancer portion into a set of sequence elements that each comprise a k-mer fragment of the enhancer portion of the polynucleotide sequence;
d) processing the set of sequence elements with a machine learning model to determine a transcriptional regulatory score associated with each sequence element of the set of sequence elements;
e) identifying at least one sequence element of the set of sequence elements comprising a trough of at least 20 nucleotides between peaks with a transcriptional regulatory score below a threshold as a non-regulatory sequence element;
f) designing a modified polynucleotide molecule based on the determined transcriptional regulatory scores and/or identified non-regulatory sequence elements; wherein at least 10% of the nucleotides within the trough have been removed in the modified polynucleotide molecule; and
g) synthesizing an AAV viral vector comprising the modified polynucleotide molecule.
18 . An AAV viral vector comprising a modified polynucleotide that was synthesized using a method comprising the steps of:
a) receiving a set of genetic data comprising a regulatory polynucleotide sequence, optionally having a known transcriptional regulatory activity;
b) identifying an enhancer portion of the polynucleotide sequence using a machine learning model based on k-mers;
c) processing the enhancer portion into a set of sequence elements that each comprise a k-mer fragment of the enhancer portion of the polynucleotide sequence;
d) processing the set of sequence elements with a machine learning model to determine a transcriptional regulatory score associated with each sequence element of the set of sequence elements;
e) identifying at least one sequence element of the set of sequence elements comprising a trough of at least 20 nucleotides between peaks with a transcriptional regulatory score below a threshold as a non-regulatory sequence element;
f) designing a modified polynucleotide molecule based on the determined transcriptional regulatory scores and/or identified non-regulatory sequence elements; wherein at least 10% of the nucleotides within the trough have been removed in the modified polynucleotide molecule; and
g) synthesizing an AAV viral vector comprising the modified polynucleotide molecule.