Prediction of phenotypes using recommender systems

Abstract

A computing server may use one or more recommender systems to predict phenotypes of individuals based on survey responses, other phenotypes, environmental factors, and genetic data of the individuals. The computing server may retrieve survey responses of a set of individuals regarding a set of phenotypes of the individuals. The computing server may construct a matrix that includes the values of the phenotypes. The computing server may predict an undetermined phenotype of a target individual using collaborative filtering, which provides the prediction based on other phenotypes of the target individuals and based on at least another individual's phenotypes. The computing server may also predict a target phenotype based on the phenotype of other individuals who are similar to the target individual. The computing server may determine another individual is similar to the target individual based on the length of identity-by-descent (IBD) segments between the two individuals.

What is claimed is: 
     
         1 . A computer-implemented method, comprising:
 retrieving user data of a set of individuals, the set of individuals comprising a target individual, the user data including at least a genetic dataset and a phenotype dataset;   converting a subset of values of the user data into a set of feature vectors, a feature vector in the set of feature vectors corresponding to an individual and comprising one or more numerical representations of genetic data related to the individual and one or more numerical representations of the phenotype data related to the individual;   clustering the set of feature vectors into a plurality of clusters, each cluster comprising one or more feature vectors representing one or more individuals;   identifying one or more similar individuals who are similar to the target individual, the similar individuals belonging to one of the clusters to which the target individual belongs; and   predicting values of one or more phenotypes of the target individual based on values of corresponding one or more phenotypes of the similar individuals.   
     
     
         2 . The computer-implemented method of  claim 1 , wherein at least one feature in the set of feature vectors is determined based on a length of identity-by-descent (IBD) segments shared between two individuals. 
     
     
         3 . The computer-implemented method of  claim 2 , wherein the similar individuals belong to an IBD genetic community to which the target individual belongs, the IBD genetic community determined by a Louvain method. 
     
     
         4 . The computer-implemented method of  claim 1 , wherein at least one feature in the set of feature vectors is determined based on an ethnicity composition of an individual. 
     
     
         5 . The computer-implemented method of  claim 1 , further comprising:
 retrieving literature data describing phenotype relationships, wherein at least one feature in the set of feature vectors is determined based on one or more relationships between a particular phenotype and other phenotypes as indicated in the literature data.   
     
     
         6 . The computer-implemented method of  claim 1 , further comprising:
 predicting survey responses of one or more other individuals other than the target individual, each of the other individuals associated with a set of other similar individuals, the set of other similar individuals and the other individual belong to a particular cluster, the survey responses of the other individual predicted based on survey responses of the set of other similar individuals;   comparing the predicted survey responses to actual survey responses to determine reliability of the actual survey responses; and   responsive to a survey question whose responses have the reliability that is lower than a threshold, adjusting or removing the survey question.   
     
     
         7 . The computer-implemented method of  claim 6 , wherein the survey responses comprising responses to one or more survey questions, the survey questions comprising one or more of the following: an appearance trait question, a social-economical question, a cultural question, a preference question, a geographical question, a health-related question, or a family health history question. 
     
     
         8 . The computer-implemented method of  claim 1 , further comprising:
 adjusting a predicted value of one of the one or more phenotypes of the target individual based on other phenotypes of the target individual.   
     
     
         9 . The computer-implemented method of  claim 1 , further comprising:
 retrieving survey responses of the set of individuals;   constructing one or more matrices that arrange the set of individuals in a first dimension of the matrices and the survey responses in a second dimension of the matrices;   identifying a plurality of undetermined survey responses that represent sparsity of the matrices;   determining, based on the plurality of the clusters, predictions of one or more of the survey responses for a first subset of individuals whose corresponding matrices having the sparsity higher than a threshold, each predicted survey response for each particular individual in the first subset of individuals determined based on survey responses of other individuals who belong to a cluster to which the particular individual belongs; and   determining, based on collaborative filtering, predictions of one or more of the survey responses for a second subset of individuals whose corresponding matrices having the sparsity lower than a threshold.   
     
     
         10 . The computer-implemented method of  claim 1 , wherein the plurality of clusters are clustered by one or more clustering techniques comprising K-Means clustering, mean-shift clustering, hierarchical clustering, or Louvain method. 
     
     
         11 . The computer-implemented method of  claim 1 , wherein the plurality of clusters are clustered based on Euclidean distances or cosine similarities among the set of feature vectors. 
     
     
         12 . The computer-implemented method of  claim 1 , wherein one of the phenotypes of the target individual predicted is an appearance trait, a medical trait, a wellness trait, or a preference of the target individual. 
     
     
         13 . A non-transitory computer readable medium configured to store computer code comprising instructions, the instructions, when executed by one or more processors, cause the one or more processors to perform steps comprising:
 retrieving user data of a set of individuals, the set of individuals comprising a target individual, the user data including at least a genetic dataset and a phenotype dataset;   converting a subset of values of the user data into a set of feature vectors, a feature vector in the set of feature vectors corresponding to an individual and comprising one or more numerical representations of genetic data related to the individual and one or more numerical representations of the phenotype data related to the individual;   clustering the set of feature vectors into a plurality of clusters, each cluster comprising one or more feature vectors representing one or more individuals;   identifying one or more similar individuals who are similar to the target individual, the similar individuals belonging to one of the clusters to which the target individual belongs; and   predicting values of one or more phenotypes of the target individual based on values of corresponding one or more phenotypes of the similar individuals.   
     
     
         14 . The non-transitory computer readable medium of  claim 13 , wherein the steps further comprise:
 predicting survey responses of one or more other individuals other than the target individual, each of the other individuals associated with a set of other similar individuals, the set of other similar individuals and the other individual belong to a particular cluster, the survey responses of the other individual predicted based on survey responses of the set of other similar individuals;   comparing the predicted survey responses to actual survey responses to determine reliability of the actual survey responses; and   responsive to a survey question whose responses have the reliability that is lower than a threshold, adjusting or removing the survey question.   
     
     
         15 . The non-transitory computer readable medium of  claim 14 , wherein the survey responses comprising responses to one or more survey questions, the survey questions comprising one or more of the following: an appearance trait question, a social-economical question, a cultural question, a preference question, a geographical question, a health-related question, or a family health history question. 
     
     
         16 . The non-transitory computer readable medium of  claim 13 , wherein the steps further comprise:
 retrieving survey responses of the set of individuals;   constructing one or more matrices that arrange the set of individuals in a first dimension of the matrices and the survey responses in a second dimension of the matrices;   identifying a plurality of undetermined survey responses that represent sparsity of the matrices;   determining, based on the plurality of the clusters, predictions of one or more of the survey responses for a first subset of individuals whose corresponding matrices having the sparsity higher than a threshold, each predicted survey response for each particular individual in the first subset of individuals determined based on survey responses of other individuals who belong to a cluster to which the particular individual belongs; and   determining, based on collaborative filtering, predictions of one or more of the survey responses for a second subset of individuals whose corresponding matrices having the sparsity lower than a threshold.   
     
     
         17 . A system comprising:
 one or more processors; and   memory configured to store computer code comprising instructions, the instructions, when executed by the one or more processors, cause the one or more processors to perform steps comprising:
 retrieving user data of a set of individuals, the set of individuals comprising a target individual, the user data including at least a genetic dataset and a phenotype dataset; 
 converting a subset of values of the user data into a set of feature vectors, a feature vector in the set of feature vectors corresponding to an individual and comprising one or more numerical representations of genetic data related to the individual and one or more numerical representations of the phenotype data related to the individual; 
 clustering the set of feature vectors into a plurality of clusters, each cluster comprising one or more feature vectors representing one or more individuals; 
 identifying one or more similar individuals who are similar to the target individual, the similar individuals belonging to one of the clusters to which the target individual belongs; and 
 predicting values of one or more phenotypes of the target individual based on values of corresponding one or more phenotypes of the similar individuals. 
   
     
     
         18 . The system of  claim 17 , wherein the steps further comprise:
 predicting survey responses of one or more other individuals other than the target individual, each of the other individuals associated with a set of other similar individuals, the set of other similar individuals and the other individual belong to a particular cluster, the survey responses of the other individual predicted based on survey responses of the set of other similar individuals;   comparing the predicted survey responses to actual survey responses to determine reliability of the actual survey responses; and   responsive to a survey question whose responses have the reliability that is lower than a threshold, adjusting or removing the survey question.   
     
     
         19 . The system of  claim 18 , wherein the survey responses comprising responses to one or more survey questions, the survey questions comprising one or more of the following: an appearance trait question, a social-economical question, a cultural question, a preference question, a geographical question, a health-related question, or a family health history question. 
     
     
         20 . The system of  claim 17 , wherein the steps further comprise:
 retrieving survey responses of the set of individuals;   constructing one or more matrices that arrange the set of individuals in a first dimension of the matrices and the survey responses in a second dimension of the matrices;   identifying a plurality of undetermined survey responses that represent sparsity of the matrices;   determining, based on the plurality of the clusters, predictions of one or more of the survey responses for a first subset of individuals whose corresponding matrices having the sparsity higher than a threshold, each predicted survey response for each particular individual in the first subset of individuals determined based on survey responses of other individuals who belong to a cluster to which the particular individual belongs; and   determining, based on collaborative filtering, predictions of one or more of the survey responses for a second subset of individuals whose corresponding matrices having the sparsity lower than a threshold.