Protein secondary structure prediction

Abstract

An artificial intelligence model receives a FTIR spectrum of a given ingredient to predict its protein secondary structure. The model includes three artificial modules, which generate three predicted values corresponding to structural categories (e.g., α-helix, β-sheet, and other) of the predicted secondary structure. Proteins may be compared for similarity based on predicted values corresponding to the structural categories of the predicted secondary structure.

What is claimed is: 
     
         1 . A method comprising:
 applying an artificial intelligence model, configured with a set of hyperparameters, to a Fourier Transform Infrared Spectroscopy (FTIR) spectrum based datum of a particular ingredient in which its secondary structure is unknown;   wherein the artificial intelligence model is trained on a plurality of raw ingredients in which their secondary structures are known, by matching FTIR spectra based data of the plurality of raw ingredients with secondary structure data of the plurality of raw ingredients to obtain the set of hyperparameters;   in response to applying the artificial intelligence model, generating protein profile data describing a predicted secondary structure of the particular ingredient.   
     
     
         2 . The method of  claim 1 , wherein the artificial intelligence model comprises a plurality of artificial intelligence components, wherein during applying the artificial intelligence model, each of the plurality of artificial intelligence components generates a particular value representing a particular aspect of the predicted secondary structure of the particular ingredient. 
     
     
         3 . The method of  claim 2 , wherein each of the plurality of artificial intelligence components comprises a plurality of sub-models, wherein each sub-model of the plurality of sub-models is applied to at least a portion of the FTIR spectrum based datum of the particular ingredient. 
     
     
         4 . The method of  claim 3 , wherein each of the plurality of artificial intelligence components further comprises a stacking ensemble regressor applied on first prediction values output from the plurality of sub-models to generate the particular value. 
     
     
         5 . The method of  claim 2 , wherein the protein profile data includes a plurality of particular values representing a plurality of particular aspects of the predicted secondary structure of the particular ingredient. 
     
     
         6 . The method of  claim 5 , wherein the plurality of particular aspects includes a first aspect relating to α-helix structural group, a second aspect relating to β-sheet structural group, and a third aspect relating to coils and turns structural group. 
     
     
         7 . The method of  claim 1 , further comprising using the predicted secondary structure of the particular ingredient to update stored secondary structure data. 
     
     
         8 . The method of  claim 7 , further comprising using the stored secondary structure data to determine a candidate ingredient that satisfies a protein replacement request, based on one or more similarity metrics. 
     
     
         9 . One or more non-transitory computer-readable storage media storing one or more instructions programmed which, when executed by one or more computing devices, cause:
 applying an artificial intelligence model, configured with a set of hyperparameters, to a Fourier Transform Infrared Spectroscopy (FTIR) spectrum based datum of a particular ingredient in which its secondary structure is unknown;   wherein the artificial intelligence model is trained on a plurality of raw ingredients in which their secondary structures are known, by matching FTIR spectra based data of the plurality of raw ingredients with secondary structure data of the plurality of raw ingredients to obtain the set of hyperparameters;   in response to applying the artificial intelligence model, generating protein profile data describing a predicted secondary structure of the particular ingredient.   
     
     
         10 . The one or more non-transitory computer-readable storage media of  claim 9 , wherein the artificial intelligence model comprises a plurality of artificial intelligence components, wherein during applying the artificial intelligence model, each of the plurality of artificial intelligence components generates a particular value representing a particular aspect of the predicted secondary structure of the particular ingredient. 
     
     
         11 . The one or more non-transitory computer-readable storage media of  claim 10 , wherein each of the plurality of artificial intelligence components comprises a plurality of sub-models, wherein each sub-model of the plurality of sub-models is applied to at least a portion of the FTIR spectrum based datum of the particular ingredient. 
     
     
         12 . The one or more non-transitory computer-readable storage media of  claim 11 , wherein each of the plurality of artificial intelligence components further comprises a stacking ensemble regressor applied on first prediction values output from the plurality of sub-models to generate the particular value. 
     
     
         13 . The one or more non-transitory computer-readable storage media of  claim 10 , wherein the protein profile data includes a plurality of particular values representing a plurality of particular aspects of the predicted secondary structure of the particular ingredient. 
     
     
         14 . The one or more non-transitory computer-readable storage media of  claim 13 , wherein the plurality of particular aspects includes a first aspect relating to α-helix structural group, a second aspect relating to R-sheet structural group, and a third aspect relating to coils and turns structural group. 
     
     
         15 . A computing system comprising:
 one or more computer systems comprising one or more hardware processors and storage media; and   instructions stored in the storage media and which, when executed by the computing system, cause the computing system to perform:
 applying an artificial intelligence model, configured with a set of hyperparameters, to a Fourier Transform Infrared Spectroscopy (FTIR) spectrum based datum of a particular ingredient in which its secondary structure is unknown; 
 wherein the artificial intelligence model is trained on a plurality of raw ingredients in which their secondary structures are known, by matching FTIR spectra based data of the plurality of raw ingredients with secondary structure data of the plurality of raw ingredients to obtain the set of hyperparameters; 
 in response to applying the artificial intelligence model, generating protein profile data describing a predicted secondary structure of the particular ingredient. 
   
     
     
         16 . The computing system of  claim 15 , wherein the artificial intelligence model comprises a plurality of artificial intelligence components, wherein during applying the artificial intelligence model, each of the plurality of artificial intelligence components generates a particular value representing a particular aspect of the predicted secondary structure of the particular ingredient. 
     
     
         17 . The computing system of  claim 16 , wherein each of the plurality of artificial intelligence components comprises a plurality of sub-models, wherein each sub-model of the plurality of sub-models is applied to at least a portion of the FTIR spectrum based datum of the particular ingredient. 
     
     
         18 . The computing system of  claim 17 , wherein each of the plurality of artificial intelligence components further comprises a stacking ensemble regressor applied on first prediction values output from the plurality of sub-models to generate the particular value. 
     
     
         19 . The computing system of  claim 16 , wherein the protein profile data includes a plurality of particular values representing a plurality of particular aspects of the predicted secondary structure of the particular ingredient. 
     
     
         20 . The computing system of  claim 19 , wherein the plurality of particular aspects includes a first aspect relating to α-helix structural group, a second aspect relating to β-sheet structural group, and a third aspect relating to coils and turns structural group.