Back to results
US 2026/0185080 A1Published application
Methods, systems, and compositions of generating and analyzing polypeptide libraries
AssigneeProtillion Biosciences, Inc.
InventorsCurtis James LAYTON, Pavanapuresan Pushpagiri VAIDYANATHAN, Michael Roy GOTRIK
Life of this patent
Estimating the term…
Abstract
Methods, systems, and composition for the analysis of polypeptides and the generation of polypeptide libraries are disclosed. Analysis of polypeptide library may be used to generate polypeptide with particular characteristics. Antibodies with high affinities may be generated using the methods, systems and compositions disclosed.
Claims28
A computer-implemented method of optimizing training of a predictive algorithm, comprising:
- (a) providing a first library of polynucleotides encoding a first library of variant polypeptides;
- (b) processing said first library of polynucleotides to produce said first library of variant polypeptides, wherein said variant polypeptides are attached to said first library of polynucleotides;
- (c) identifying one or more characteristics comprising an equilibrium binding constant, a kinetic binding constant, a protein stability measurement, an enzyme activity, a fractional activity, a nonspecific binding potential, an aggregation potential, a hydrophobicity, a protein expression level, or a maturation time of at least a portion of said first library of variant polypeptides;
- (d) providing a second library of polynucleotides encoding a second library of variant polypeptides selected based at least on said one or more characteristics identified in (c);
- (e) processing said second library of polynucleotides to produce said second library of variant polypeptides wherein said variant polypeptides are attached to said second library of polynucleotides;
- and
- (f) creating a training set comprising said second library of variant polypeptides, wherein a combinatorial sequence space of said second library of variant polypeptides is less than about 10%, less than about 1%, or less than about 0.1% of a combinatorial sequence space of said first library of variant polypeptides.
4 more elements
A computer-implemented method of optimizing training of a predictive algorithm, comprising:
- (a) providing a first library of polynucleotides encoding a first library of variant polypeptides;
- (b) processing said first library of polynucleotides to produce said first library of variant polypeptides wherein said variant polypeptides are attached to said first library of polynucleotides;
- (c) identifying one or more characteristics comprising an equilibrium binding constant, a kinetic binding constant, a protein stability measurement, an enzyme activity, a fractional activity, a nonspecific binding potential, or an aggregation potential, a hydrophobicity, a protein expression level, or a maturation time of at least a portion of said first library of variant polypeptides;
- (d) providing a second library of polynucleotides encoding a second library of variant polypeptides selected based at least on said one or more characteristics identified in (c);
- (e) processing said second library of polynucleotides to produce said second library of variant polypeptides, wherein said variant polypeptides are attached to said second library of polynucleotides;
- and
- (f) creating a training set comprising said second library of variant polypeptides, wherein said second library of variant polypeptides comprises a polypeptide with an at least 10% improvement in said one or more characteristics identified in (c) compared to a reference polypeptide of said first library of variant polypeptides.
4 more elements
A computer-implemented method of optimizing training of a predictive algorithm, comprising:
- (a) providing a library of variant polypeptides, wherein said library of variant polypeptides comprise a plurality of variations of a starting polypeptide;
- (b) identifying one or more characteristics comprising an equilibrium binding constant, a kinetic binding constant, a protein stability measurement, an enzyme activity, a fractional activity, a nonspecific binding potential, an aggregation potential, a hydrophobicity, a protein expression level, or a maturation time of at least a portion of said library of variant polypeptides;
- and
- (c) creating a training set comprising a at least a portion of said library of variant polypeptides selected based at least on one or more characteristics identified in (b), wherein a combinatorial sequence space of said portion of said library of variant polypeptides is less than about 10%, less than about 1%, or less than about 0.1% of a combinatorial sequence space of said library of variant polypeptides.
1 more element
Dependencies are inferred from the claim text.