Compressed Sensing for Simultaneous Measurement of Multiple Different Biological Molecule Types in a Sample

Abstract

A method and apparatus for simultaneously determining multiple different biological molecule types in a sample include labeling each different biological molecule type in a biological sample with a unique combination of a plurality of labels. Each different biological molecule type is selected from a population of M different biological molecules types. The plurality of labels is selected from a population of L different labels; and, M is greater than L. Measurements are obtained of relative abundances of the L different labels in the sample. Relative abundance of up to M different biological molecule types in the sample are determined based on the measurements and a method of compressed sensing.

What is claimed is: 
     
         1 . A method comprising:
 labeling each different biological molecule type in a biological sample with a unique combination of a plurality of labels, wherein
 each different biological molecule type is selected from a population of M different biological molecules types, 
 the plurality of labels is selected from a population of L different labels, and 
 M is greater than L; 
   obtaining measurements of relative abundances of the L different labels in the sample; and   determining relative abundance of up to M different biological molecule types in the sample based on the measurements and a method of compressed sensing.   
     
     
         2 . A method as recited in  claim 1 , wherein:
 the plurality of labels is selected from a population of about different fluorophores with corresponding different fluorescent wavelengths; and   obtaining measurements of relative abundance of the L different labels in the sample further comprises obtaining measurements of intensity of each of the corresponding fluorescent wavelengths.   
     
     
         3 . A method as recited in  claim 1 , wherein:
 the plurality of labels is selected from a population of different isotopes of elements which are not otherwise found in biological organisms and which have ions of different corresponding masses; and   obtaining measurements of relative abundance of the L different labels in the sample further comprises reducing the sample to ions of its constituent elements and obtaining measurements of abundance of each of the corresponding masses in a time of flight mass spectrometer.   
     
     
         4 . A method as recited in  claim 1 , wherein:
 each different biological molecule type is a different protein; and   labeling each different biological molecule type in the biological sample with a unique combination of the plurality of labels further comprises
 attaching each label in the unique combination to an antibody that binds to the different protein to produce a plurality of labeled antibodies, 
 contacting the biological sample with the plurality of labeled antibodies for each different protein. 
   
     
     
         5 . A method as recited in  claim 4 , wherein the biological sample comprises whole cells and each different protein is a different receptor for an outer cell membrane of at least some cells. 
     
     
         6 . A method as recited in  claim 4 , wherein the biological sample comprises permeabilized cells and each different protein is found inside at least some cells. 
     
     
         7 . A method as recited in  claim 1 , wherein the method of compressed sensing comprises selecting abundances for up to M different molecular types by minimizing an order one distance metric designated l 1  between computed and measured abundances of the L different labels. 
     
     
         8 . A method as recited in  claim 1 , wherein the method of compressed sensing comprises selecting abundances for up to M different molecular types by minimizing a sum of a first term and an order one distance metric designated l 1  of computed abundances of the L different labels, wherein the first term is a parameter λ times a computed abundance of the biological molecule. 
     
     
         9 . A method as recited in  claim 8 , wherein a value for the parameter λ is chosen to provide acceptable performance for a test set with known abundances of the different biological molecule types. 
     
     
         10 . A method as recited in  claim 1 , wherein the unique combination of labels for each different molecule type is assigned at random. 
     
     
         11 . A method as recited in  claim 1 , wherein the unique combination of labels for each different molecule type is assigned at random with up to y of the L labels, wherein y is much less than L. 
     
     
         12 . A method as recited in  claim 1 , wherein the unique combination of labels for each different molecule type is chosen to provide acceptable performance for a test set with known abundances of the different biological molecule types. 
     
     
         13 . A method as recited in  claim 1 , further comprising obtaining the biological sample such that the biological sample includes a number N of the biological molecule types, wherein N is less than about L. 
     
     
         14 . An apparatus comprising:
 at least one processor; and   at least one memory including one or more sequences of instructions,   the at least one memory and the one or more sequences of instructions configured to, with the at least one processor, cause the apparatus to perform at least the following,
 determine a set of L different labels for which abundance can be measured; 
 determine a set of M different biological molecule types to be detected; 
 determine an assignment matrix that indicates a unique combination of different labels for each molecule type; 
 obtain measurements of abundance of labels in a biological sample; and 
 determine abundances for up to M different molecule types in the biological sample based on the measurements and techniques of compressed sensing. 
   
     
     
         15 . An apparatus as recited in  claim 14 , wherein to determine abundances for up to M different molecule types in the biological sample based on techniques of compressed sensing further comprises to determine abundances for up to M different molecular types by minimizing a sum of a first term and an order one distance metric designated l 1  of computed abundances of the L different labels, wherein the first term is a parameter λ times a computed abundance of the biological molecule 
     
     
         16 . An apparatus as recited in  claim 15 , wherein the at least one memory and the one or more sequences of instructions, with the at least one processor, are further configured to cause the apparatus to determine a value for the parameter λ that provides acceptable performance for a test set with known abundances of the different biological molecule types. 
     
     
         17 . An apparatus as recited in  claim 14 , wherein to determine the assignment matrix that indicates the unique combination of different labels for each molecule type further comprises to determine the assignment matrix that provides acceptable performance for a test set with known abundances of the different biological molecule types. 
     
     
         18 . An apparatus as recited in  claim 14 , further comprising a cell sorter, wherein the cell sorter is operated based on the abundances determined for the up to M different molecule types in the biological sample. 
     
     
         19 . An apparatus comprising:
 means for labeling each different biological molecule type in a biological sample with a unique combination of a plurality of labels, wherein
 each different biological molecule type is selected from a population of M different biological molecules types, 
 the plurality of labels is selected from a population of L different labels, and 
 M is greater than L; 
   means for obtaining measurements of relative abundance of the L different labels in the sample; and   means for determining relative abundance of up to M different molecule types in the sample based on the measurements and a method of compressed sensing.   
     
     
         20 . A non-transient computer-readable medium carrying one or more sequences of instructions, wherein execution of the one or more sequences of instructions by one or more processors causes the one or more processors to perform to:
 determine a set of L different labels for which abundance can be measured;   determine a set of M different biological molecule types to be detected;   determine an assignment matrix that indicates a unique combination of different labels for each molecule type;   obtain measurements of abundance of labels in a biological sample; and   determine abundances for up to M different molecule types in the biological sample based on the measurements and techniques of compressed sensing.   
     
     
         21 . A kit comprising a plurality of reagents for labeling each different biological molecule type with a unique combination of a plurality of labels, wherein:
 each different biological molecule type is selected from a population of M different biological molecules types;   the plurality of labels is selected from a population of L different labels that are distinguishable in a measurement apparatus; and   M is greater than L.   
     
     
         22 . A method as recited in  claim 1 , wherein labeling each different biological molecule type in the biological sample with a unique combination of a plurality of labels further comprises equalizing ratios of each label in each unique combination. 
     
     
         23 . A method as recited in  claim 1 , further comprising
 labeling each different biological molecule type in a separate subsample of the sample with a second unique combination of a plurality of labels selected from a population of M different labels;   obtaining second measurements of relative abundances of the M different labels in the sample; and   determining second relative abundance of the M different biological molecule types in the sample based on the second measurements; and   determining adequacy of the unique combination of the plurality of labels selected from the population of L different labels based on comparing the second relative abundances to the relative abundances based on compressed sensing.   
     
     
         24 . A method as recited in  claim 23 , further comprising repeating all steps for N new biological molecule types with a third unique combination of the labels selected from a population of L different labels wherein N is greater than L and a fourth unique combination of a plurality of labels selected from a population of N different labels.