Systems and Methods for SNP Characterization and Identifying off Target Variants

Abstract

Methods for processing data using information gained from examining biological materials identifies and characterized probes for Single Nucleotide Polymorphisms and identifies Off Target Variants.

1 . A method of genotyping polymorphisms using a computer system comprising:
 accessing genotype clustering data from a plurality of polymorphism (e.g., SNPs and indels) probeset locations from a plurality of samples, the genotype clustering data comprising genotype cluster calls of samples, cluster centers, and variance;   accessing polymorphisms probeset sample data that indicates contrast and intensity detected at a probeset location across a number of samples;   determining one or more quality metrics for each probeset;   using said quality metrics and said computer system, choosing a selected probeset from two or more probesets available for a polymorphisms; and   outputting the selected probeset.   
     
     
         2 . The method according to  claim 1  further wherein said quality metrics comprise:
 a heterozygote strength offset (optionally referred to as HetSO) metric determined by a distance from a center of a heterozygote cluster to a line connecting centers of two homozygote clusters where larger distances in the direction of decreasing contrast are associated with or indicate off target variant (optionally referred to as OTV) clusters and where said heterozygote cluster and said two homozygote clusters are plotted with contrast as one dimension and intensity as another dimension. 
 
     
     
         3 . The method according to  claim 1  further wherein said quality metrics comprise:
 a homozygote ratio offset (optionally referred to as HomRO) determined by a location, in a contrast dimension, of a homozygous genotype cluster center that is closest to zero. 
 
     
     
         4 . The method according to  claim 3  further wherein, for populated homozygote clusters AA and BB:
 if both of the populated homozygote clusters AA and BB are on the proper side of zero contrast, the HomRO=min(X AA , abs(X BB )), where X AA =a posterior mean of the AA cluster, and X BB =a posterior mean of the BB cluster; 
 if both are positive (on the AA cluster side of zero), HomRO=−X BB , wherein negative designation indicates that BB is on a wrong side of 0; 
 if both are negative (on the BB side of zero), HomRO=X AA . 
 
     
     
         5 . The method according to  claim 1  further comprising:
 classifying each probeset as into one of a plurality of probeset classifications; and 
 labeling data representing said probesets according to said classifications to provide for down stream analysis and visualizations according to said classifications. 
 
     
     
         6 . (canceled) 
     
     
         7 . The method according to  claim 5  further comprising:
 determining an order of priority for said classifications; 
 selecting a probeset with a highest priority. 
 
     
     
         8 . The method according to  claim 7  further comprising:
 breaking ties when two or more probesets are in a highest priority by choosing a probeset with a highest quality metric. 
 
     
     
         9 . The method according to  claim 1  further comprising:
 using said selected probeset to genotype a plurality of samples of the polymorphism. 
 
     
     
         10 . The method according to  claim 1  further comprising:
 using said selected probeset to manufacture probes for a genomic array. 
 
     
     
         11 . The method according to  claim 1  further comprising:
 using said method during a diagnostic genomic assay in order to select which probeset will be used for making genotype calls for a polymorphism. 
 
     
     
         12 . The method according to  claim 1  further comprising:
 using said method during a screening genomic assay in order to select which probesets will be manufactured for use on a genotype assay device or system. 
 
     
     
         13 - 15 . (canceled) 
     
     
         16 . A method of genotyping single nucleotide polymorphisms (SNPs) by identifying off-target variants (OTVs) using a computer system comprising:
 accessing genotype clustering data from a number of samples from a probeset location, the genotype clustering data comprising genotype cluster calls of samples; cluster centers, and variance;   accessing sample data that indicates contrast and intensity detected at a probeset location across a number of samples;   searching for optimal clustering of said sample data in both contrast and intensity dimensions, allowing for an OTV genotype cluster;   iteratively updating sample assignments to clusters and cluster centers until convergence; and   outputting genotyping cluster results for said samples.   
     
     
         17 . The method according to  claim 16  further wherein said identifying a probeset as an OTV probeset is not based on characteristics of any single sample. 
     
     
         18 . (canceled) 
     
     
         19 . The method according to  claim 16  further wherein said searching and said iteratively updating use portions of sample intensity data that were not used to generate said genotype clustering data. 
     
     
         20 . The method according to  claim 16  further wherein said searching and said iteratively updating give sample intensity data more weight than was used for clustering said genotype clustering data. 
     
     
         21 . The method according to  claim 16  further comprising:
 determining a distance from the center of a heterozygote cluster to a line connecting homozygote clusters when said samples are plotted according to contrast and intensity; 
 comparing the distance to a threshold; 
 label as off target variants, clusters characterized by good cluster resolution according to a plurality of cluster parameters and by said distance exceeding a threshold. 
 
     
     
         22 . (canceled) 
     
     
         23 . The method according to  claim 16  further comprising:
 determine if samples outside of a populated OTV cluster constitute a genotype. 
 
     
     
         24 . The method according to  claim 16  further wherein:
 the genotype clustering data is generated by an automated clustering algorithm comprising one or more selected from the group: AxiomGT1, BRLMM-P, or any other suitable clustering algorithm. 
 
     
     
         25 . The method according to  claim 16  further wherein:
 said searching and iteratively updating is performed by a clustering algorithm selected from the group: Expectation-Maximization (EM), 2D Baum-Welch, Viterbi, HMM analysis. 
 
     
     
         26 . (canceled) 
     
     
         27 . The method according to  claim 16  further comprising:
 determining a cluster is an OTV cluster when: 
 the samples overall have good genotype cluster properties with the exception of a cluster in an OTV position; 
 a Fishers linear discriminant (FLD) of the cluster is greater than or equal to a threshold; 
 a homozygote ratio offset (optionally referred to as HomRO) determined by a location, in a contrast dimension, of a homozygous genotype cluster center that is closest to zero is greater than or equal to a threshold; 
 the homozygote cluster or clusters is offset from zero in the contrast dimension; and 
 the distance from the center of a heterozygote cluster to a line connecting homozygote clusters is less than or equal to a threshold. 
 
     
     
         28 . (canceled) 
     
     
         29 . The method according to  claim 16  further comprising:
 selecting between four and three genotype clusters models by, for each model: 
 initiating by using posterior information of genotype cluster from a clustering algorithm; 
 assigning each sample into one of genotype clusters according to expectation/likelihood, appropriate for the model, e.g., AA, AB, BB, and OTV for the four cluster model and AA, BB, and OTV for the three cluster model; 
 calculating the genotype cluster center locations by maximizing likelihood; and 
 selecting the model with the greatest maximum likelihood. 
 
     
     
         30 . (canceled) 
     
     
         31 . The method according to  claim 16  further comprising:
 using said selected probeset to genotype said SNP. 
 
     
     
         32 . The method according to  claim 16  further comprising:
 using said selected probeset to select probes for use in a genomic assay system or device, such as a genotyping array. 
 
     
     
         33 - 34 . (canceled) 
     
     
         35 . A method using a computer system to select a preferred SNP probeset from a plurality of SNP probesets comprising:
 determining a plurality of first SNP probeset quality control metrics that characterize genotype clusters for an SNP probeset;   classify each probeset for an SNP using the first and second SNP probeset quality control metrics;   select a probeset with the highest priority classification.   
     
     
         36 . The method according to  claim 35  further wherein:
 the SNP probeset classifications are two or more selected from the group consisting of: 
 (1) clusters meeting threshold criteria for good cluster resolution and having at least 2 samples of the minor allele optionally referred to as “PolyHighResolution”; 
 (2) clusters characterized by less than 2 samples of the minor allele, optionally referred to as “MonoHighResolution”; 
 (3) probesets with an a populated cluster of samples indicating the presence of one or more off target variants in the probeset samples optionally referred to as “OTV;” 
 (4) two clusters with no examples of the minor homozygous genotypes, optionally referred to as “NoMinorHom;” and 
 (5) a genotype call rate for the probeset below threshold, with other cluster properties above threshold, optionally referred to as “Call Rate Below Threshold.” 
 
     
     
         37 - 38 . (canceled) 
     
     
         39 . The method according to  claim 35  further comprising:
 using said selected probeset to genotype said SNP. 
 
     
     
         40 . The method according to  claim 35  further comprising:
 using said selected probeset to select probes for use in a genomic assay system or device, such as a genotyping array. 
 
     
     
         41 . (canceled) 
     
     
         42 . The method according to  claim 39  further comprising:
 obtaining a nucleic acid sample; 
 amplifying fragments of DNA in the nucleic acid sample; 
 hybridizing the amplified sample to an array of oligonucleotides, the array comprising:
 a plurality of different nucleotide polymorphisms probesets attached to a solid support, 
 wherein each probeset comprises one or more probes comprising 20 to 50 contiguous nucleotides from a different sequence 
 wherein each different probe is attached to a solid support in a known or determinable location of the array, 
 
 using a computer system to compare results of a plurality of probesets targeting a single SNP; 
 using the computer system to automatically select one or more selected probesets from said plurality of probesets using probeset selection criteria 
 determining the genotype of the sample for a plurality of each of the single nucleotide polymorphisms by using the one or more selected probesets from that probeset. 
 
     
     
         43 . The method of  claim 42  wherein the solid support is a plurality of beads wherein a plurality of each probe is attached to a different bead. 
     
     
         44 . The method of  claim 42  wherein the step of amplifying fragments of DNA in the nucleic acid sample comprises fragmenting the sample with at least one restriction enzyme, ligating an adapter to at least some of the fragments to generate adapter ligated fragments, and amplifying at least some of the adapter ligated fragments by polymerase chain reaction using a primer to the adapter. 
     
     
         45 . (canceled) 
     
     
         46 . A computer program stored on a non-transitory computer readable medium, the computer program comprising code to perform:
 accessing genotyping data comprising intensity and contrast data for one or more probesets for a polymorphism;   determining a plurality of probeset quality control metrics for said one or more probesets;   when said genotyping data includes two or more of said probesets of said polymorphism, selecting at least one of said two or more probesets for providing an improved genotyping result for said polymorphism;   outputting data indicating said quality control metrics or said at least one probeset or said genotyping result or any combination thereof.   
     
     
         47 - 50 . (canceled) 
     
     
         51 . The computer program of  claim 46 , further comprising code to perform:
 classifying said two or more probesets as containing or not containing an off target variant cluster based on said metrics.   
     
     
         52 . The computer program of  claim 46 , wherein said accessing further comprises accessing prior genotype clustering results for said two or more probesets. 
     
     
         53 . The computer program of  claim 52 , wherein said prior genotype clustering results comprise genotype cluster assignment, cluster centers, and cluster variances as produced by Affymetrix Axiom Genotyping Arrays. 
     
     
         54 - 55 . (canceled)