Methods for cell label classification

Abstract

Disclosed herein are methods and systems for classifying cell labels, for example identifying a signal cell label. In some embodiments, the method comprises: obtaining sequencing data of barcoded targets created using targets in cells barcoded using barcodes, wherein a barcode comprises a cell label and a molecular label. After ranking the cell labels, a minimum of a second derivative plot of a cumulative sum plot can be determined. Using the methods, a cell label can be classified as a signal cell label or a noise cell label based on the number of molecular labels with distinct sequences associated with the cell label and a cell label threshold.

1 .- 65 . (canceled) 
     
     
         66 . A method for identifying a signal cell label, comprising:
 (a) barcoding a plurality of targets in a plurality of cells using a plurality of barcodes to create a plurality of barcoded targets, wherein each of the plurality of barcodes comprises a cell label and a molecular label, wherein barcoded targets created from targets of different cells of the plurality of cells have different cell labels, and wherein barcoded targets created from targets of the same cell of the plurality of cells have different molecular labels;   (b) obtaining sequencing data of the plurality of barcoded targets;   (c) determining a feature vector of each cell label of the plurality of barcoded targets, wherein the feature vector comprise numbers of molecular labels with distinct sequences associated with the each cell label;   (d) determining a cluster for the each cell label of the plurality of barcoded targets based on the feature vector; and   (e) identifying the each cell label of the plurality of barcoded targets as a signal cell label or a noise cell label based on a number of cell labels in the cluster and a cluster size threshold.   
     
     
         67 .- 80 . (canceled) 
     
     
         81 . A method for identifying a signal cell label, comprising:
 (a) obtaining sequencing data of a plurality of barcoded targets, wherein the plurality of barcoded targets is create from a plurality of targets in a plurality of cells that are barcoded using a plurality of barcodes, wherein each of the plurality of barcodes comprises a cell label and a molecular label, wherein barcoded targets created from targets of different cells of the plurality of cells have different cell labels, and wherein barcoded targets created from targets of the same cell of the plurality of cells have different molecular labels;   (b) determining a feature vector of each cell label of the plurality of barcoded targets, wherein the feature vector comprise numbers of molecular labels with distinct sequences associated with the each cell label;   (c) determining a cluster for the each cell label of the plurality of barcoded targets based on the feature vector; and   (d) identifying the each cell label of the plurality of barcoded targets as a signal cell label or a noise cell label based on a number of cell labels in the cluster and a cluster size threshold.   
     
     
         82 . The method of  claim 81 , wherein determining the cluster for the each cell label of the plurality of barcoded targets based on the feature vector comprises clustering the each cell label of the plurality of barcoded targets into the cluster based on a distance of the feature vector to the cluster in a feature vector space. 
     
     
         83 . The method of  claim 81 , wherein determining the cluster for the each cell label of the plurality of barcoded targets based on the feature vector comprises: 
       projecting the feature vector from a feature vector space into a lower dimensional space; and 
       clustering the each cell label into the cluster based on a distance of the feature vector to the cluster in the lower dimensional space. 
     
     
         84 . The method of  claim 83 , wherein the lower dimensional space is a two dimensional space. 
     
     
         85 . The method of  claim 83 , wherein projecting the feature vector from the feature vector space into the lower dimensional space comprises projecting the feature vector from the feature vector space into the lower dimensional space using a t-distributed stochastic neighbor embedding (tSNE) method. 
     
     
         86 . The method of  claim 83 , wherein clustering the each cell label into the cluster based on the distance of the feature vector to the cluster in the lower dimensional space comprises clustering the each cell label into the cluster based on the distance of the feature vector to the cluster in the lower dimensional space using a density-based method. 
     
     
         87 . The method of  claim 86 , wherein the density-based method comprises a density-based spatial clustering of applications with noise (DBSCAN) method. 
     
     
         88 . The method of  claim 83 , wherein the cell label is identified as the signal cell label if the number of cell labels in the cluster is below the cluster size threshold. 
     
     
         89 . The method of  claim 83 , wherein the cell label is identified as a noise cell label if the number of cell labels in the cluster is not below the cluster size threshold. 
     
     
         90 . The method of  claim 83 , comprising determining the cluster size threshold based on the number of cell labels of the plurality of barcoded targets. 
     
     
         91 . The method of  claim 90 , wherein the cluster size threshold is a percentage of the number of cell labels of the plurality of barcoded targets. 
     
     
         92 . The method of  claim 83 , comprising determining the cluster size threshold based on the number of cell labels of the plurality of barcoded targets. 
     
     
         93 . The method of  claim 92 , wherein the cluster size threshold is a percentage of the number of cell labels of the plurality of barcoded targets. 
     
     
         94 . The method of  claim 83 , comprising determining the cluster size threshold based on numbers of molecular labels with distinct sequences associated with each cell label of the plurality of barcoded targets. 
     
     
         95 . The method of  claim 83 , comprising:
 (e) for one or more of the plurality of targets:
 1counting the number of molecular labels with distinct sequences associated with the target in the sequencing data; and 
 2) estimating the number of the target based on the number of molecular labels with distinct sequences associated with the target in the sequencing data counted in (1). 
   
     
     
         96 . A method for identifying a signal cell label, comprising:
 (a) obtaining sequencing data of a plurality of first targets of cells, wherein each first target is associated with a number of molecular labels with distinct sequences associated with each cell label of a plurality of cell labels;   (b) identifying each of the cell labels as a signal cell label or a noise cell label based on the number of molecular labels with distinct sequences associated with each of the cell labels and an identification threshold; and   (c) re-identifying at least one of the plurality of cell labels as a signal cell label identified as a noise cell label in (b) or re-identifying at least one of the cell label as a noise cell label identified as a signal cell label in (b).   
     
     
         97 .- 104 . (canceled) 
     
     
         105 . A computer system for determining the number of targets comprising:
 a hardware processor; and   non-transitory memory having instructions stored thereon, which when executed by the hardware processor causes the processor to perform the method of  claim 81 .   
     
     
         106 . A computer readable medium comprising code for performing the method for performing the method of  claim 81 .