Method, system, and product for refining an artificial intelligence model for predicting xenotransplantation compatability

Abstract

Predictive engineering of a sample derived from a genetically optimized non-human donor suitable for xenotransplantation into a human having improved quality or performance is described. A training data set is constructed from a series of libraries, including at least one library comprising genomic, proteomic, and research data specific to non-humans. A predictive machine learning model is developed based on the constructed training data set and utilized to obtain a predicted quality or performance of a plurality of sequences for a candidate sample from the non-human donor specific to a human patient or patient population. A subset of sequences is selected for evaluation from the plurality of sequences based on the predicted quality or performance and candidate samples are designed derived from the non-human donor using the selected subset of sequences.

1 - 37 . (canceled) 
     
     
         38 . A method for refining an artificial intelligence model for predicting xenotransplantation compatibility, the method comprising:
 collecting a first training data set, the first training data set comprising first experimental data from a series of libraries, the first experimental data evaluating a compatibility of xenotransplantation of a non-human sample from a non-human donor into a human recipient;   training an artificial intelligence model with the first training data set to identify one or more compatible nucleotide sequences for introduction into a genome of a non-human donor, wherein the artificial intelligence model is selected from a set of: a neural network model, a Bayesian network model, a support vector machine model, a k-nearest neighbors model, or a combination thereof;   obtaining first human nucleotide sequencing data relating to a first human patient or patient population;   identifying, using the trained artificial intelligence model and the first human nucleotide sequencing data, a first set of one or more compatible nucleotide sequences for introduction into a genome of a first non-human donor;   designing a first candidate non-human sample from the first non-human donor using the identified first set of one or more compatible nucleotide sequences;   obtaining a second training data set, the second training data set comprising second experimental data evaluating a compatibility of xenotransplantation of the first candidate non-human sample in the first human patient or patient population; and   refining the artificial intelligence model with the second training data set.   
     
     
         39 . The method of  claim 38 , wherein the first experimental data from the series of libraries comprise a major histocompatibility complex (MHC) sequence for the non-human donor and a human leukocyte antigen (HLA) sequence for the human recipient. 
     
     
         40 . The method of  claim 39 , wherein the first experimental data from the series of libraries further comprises a mixed lymphocyte reaction (MLR) assay result, wherein the MLR assay result comprises an indication of xenotransplantation compatibility of the MHC sequence with the HLA sequence. 
     
     
         41 . The method of  claim 38 , further comprising:
 modeling, in silico, an outcome of xenotransplantation of the first candidate non-human sample in the first human patient or patient population.   
     
     
         42 . The method of  claim 38 , wherein the first experimental data comprises clinical qualities or attributes specific to the first human patient or patient population comprising one or more of:
 genomic sequences, nucleotide or proteomic sequences, amino acid sequences,   HLA sequences to one or more major histocompatibility complexes (MHC) serotypes,   titers,   viability,   density,   concentration,   demographics,   results of diagnostic assays, in vitro assays, mixed lymphocyte reaction (MLR) assays,   current and past medical and family histories,   current and past medical diagnoses,   current and past clinical documentation,   current and past medications, or   observational or experimental data of a human recipient with an engineered sample.   
     
     
         43 . The method of  claim 38 , wherein the first experimental data comprises:
 protein variant data,   genomic, proteomic, and research data specific to human vertebrates, or   genomic, proteomic, and research data specific to known pathogens and diseases.   
     
     
         44 . The method of  claim 38 , wherein the second experimental data comprises a measurement of an in vitro patient outcome of xenotransplantation of the first candidate non-human sample. 
     
     
         45 . The method of  claim 38 , wherein the first set of one or more nucleotide sequences comprises one or more genomic alterations to be introduced into a genome of the first non-human donor. 
     
     
         46 . The method of  claim 45 , wherein the one or more genomic alterations comprise at least one alteration selected from the group consisting of: a single nucleotide polymorphism, nucleotide sequence insertion, nucleotide sequence deletion, nucleotide sequence replacements, or site-directed mutagenic substitution. 
     
     
         47 . The method of  claim 46 , wherein the one or more genomic alterations comprise sequences from human and non-human major histocompatibility complexes (MHC). 
     
     
         48 . The method of  claim 38 , wherein the first candidate non-human sample comprises a skin graft from the non-human donor or a nerve transplant from the non-human donor. 
     
     
         49 . The method of  claim 38 , wherein the first non-human donor is a porcine donor. 
     
     
         50 . A system for refining an artificial intelligence model for predicting xenotransplantation compatibility, the system comprising a processor and a non-transitory memory, wherein the system is configured to:
 collect a first training data set, the first training data set comprising first experimental data from a series of libraries, the first experimental data evaluating a compatibility of xenotransplantation of a non-human sample from a non-human donor into a human recipient;   train an artificial intelligence model with the first training data set to identify one or more compatible nucleotide sequences for introduction into a genome of a non-human donor, wherein the artificial intelligence model is selected from a set of: a neural network model, a Bayesian network model, a support vector machine model, a k-nearest neighbors model, or a combination thereof;   obtain first human nucleotide sequencing data relating to a first human patient or patient population;   identify, using the trained artificial intelligence model and the first human nucleotide sequencing data, a first set of one or more compatible nucleotide sequences for introduction into a genome of a first non-human donor;   design a first candidate non-human sample from the first non-human donor using the identified first set of one or more compatible nucleotide sequences;   obtain a second training data set, the second training data set comprising second experimental data evaluating a compatibility of xenotransplantation of the first candidate non-human sample in the first human patient or patient population; and   refine the artificial intelligence model with the second training data set.   
     
     
         51 . The system of  claim 50 , wherein the system is further configured to:
 model, in silico, an outcome of xenotransplantation of the first candidate non-human sample in the first human patient or patient population.   
     
     
         52 . The system of  claim 50 , wherein the second experimental data comprises a measurement of an in vitro patient outcome of xenotransplantation of the first candidate non-human sample. 
     
     
         53 . The system of  claim 50 , wherein the first set of one or more nucleotide sequences comprises one or more genomic alterations to be introduced into a genome of the first non-human donor. 
     
     
         54 . The system of  claim 53 , wherein the one or more genomic alterations comprise at least one alteration selected from the group consisting of: a single nucleotide polymorphism, nucleotide sequence insertion, nucleotide sequence deletion, nucleotide sequence replacements, or site-directed mutagenic substitution. 
     
     
         55 . The system of  claim 50 , wherein the first candidate non-human sample comprises a skin graft from the non-human donor or a nerve transplant from the non-human donor. 
     
     
         56 . The system of  claim 50 , wherein the first non-human donor is a porcine donor. 
     
     
         57 . A computer program product comprising a non-transitory computer-readable medium storing instructions thereon, wherein the instructions cause a computer to:
 collect a first training data set, the first training data set comprising first experimental data from a series of libraries, the first experimental data evaluating a compatibility of xenotransplantation of a non-human sample from a non-human donor into a human recipient;   train an artificial intelligence model with the first training data set to identify one or more compatible nucleotide sequences for introduction into a genome of a non-human donor, wherein the artificial intelligence model is selected from a set of: a neural network model, a Bayesian network model, a support vector machine model, a k-nearest neighbors model, or a combination thereof;   obtain first human nucleotide sequencing data relating to a first human patient or patient population;   identify, using the trained artificial intelligence model and the first human nucleotide sequencing data, a first set of one or more compatible nucleotide sequences for introduction into a genome of a first non-human donor;   design a first candidate non-human sample from the first non-human donor using the identified first set of one or more compatible nucleotide sequences;   obtain a second training data set, the second training data set comprising second experimental data evaluating a compatibility of xenotransplantation of the first candidate non-human sample in the first human patient or patient population; and   refine the artificial intelligence model with the second training data set.