Methods and applications of protein identification
Methods and systems for identifying a protein within a sample are provided herein. A panel of antibodies are acquired, none of which are specific for a single protein or family of proteins. Additionally, the binding properties of the antibodies in the panel are determined. Further, the protein is iteratively exposed to a panel of antibodies. Additionally, a set of antibodies which bind the protein are determined. The identity of the protein is determined using one or more deconvolution methods based on the known binding properties of the antibodies to match the set of antibodies to a sequence of a protein.
1 . A method of assaying proteins, comprising:
(a) providing a plurality of full length protein molecules immobilized on a substrate at unique spatial addresses, the unique spatial addresses being immobilized to an individual full length protein molecule of the plurality of full length protein molecules, wherein each unique spatial address is optically resolvable from each other unique spatial address on the substrate;
(b) binding composite affinity reagents to full length protein molecules of the plurality of full length protein molecules immobilized on the substrate, wherein each individual composite affinity reagent of the composite affinity reagents comprises a linker that is attached to a plurality of antibodies or fragments thereof and a common identifiable tag, and wherein each composite affinity reagent recognizes two or more differing full length the different protein molecules of the plurality of full length protein molecules;
(c) detecting signals from unique spatial addresses containing the immobilized full length protein molecules that are bound to the composite affinity reagents;
(d) repeating steps (b) and (c) using different composite affinity reagents; and
(e) detecting a binding pattern comprising presences or absences of signals from each of the different composite affinity reagents bound to the immobilized full length protein molecules at the unique spatial addresses on the substrate.
2 . The method of claim 1 , wherein step (e) further comprises determining for each of the unique spatial addresses:
(i) a set of the different composite affinity reagents that bind to the individual full length protein molecule at the unique spatial address, and
(ii) a set of the different composite affinity reagents that do not bind to the individual full length protein molecule at the unique spatial address.
3 . The method of claim 1 , wherein the substrate comprises beads.
4 . The method of claim 1 , wherein full length protein molecules of the plurality of full length protein molecules are immobilized to the substrate via nucleic acids.
5 . The method of claim 1 , wherein the signals are produced by fluorescent tags.
6 . The method of claim 1 , wherein the composite affinity reagents bind to amino acids comprising post translational modifications.
7 . The method of claim 1 , further comprising treating the plurality of full length protein molecules to remove the post translational modifications.
8 . The method of claim 1 , wherein the plurality of full length protein molecules comprises more than 1000 different full length protein molecules.
9 . The method of claim 1 , wherein the different composite affinity reagents have a known degree of non-specificity and known specificity for the full length protein molecules of the plurality of full length protein molecules.
10 . The method of claim 1 , further comprising: (f) assigning a protein identity to the full length protein molecules of the plurality of full length protein molecules based upon the binding pattern of the composite affinity reagents at the unique spatial addresses on the substrate.
11 . The method of claim 10 , wherein a number of full length protein molecules identified in step (f) is greater than the quantity of composite affinity reagents to which the plurality of full length protein molecules is exposed.
12 . The method of claim 10 , wherein the assigning of step (f) comprises executing software on a computer processor to assign a probable protein identity to each of the unique spatial addresses and a confidence for the protein identity.
13 . The method of claim 12 , wherein the software comprises a machine learning algorithm trained using binding characteristics of the composite affinity reagents as inputs.
14 . The method of claim 10 , wherein the assigning of step (f) comprises executing software on a computer processor to perform a Bayesian inference using as inputs:
(i) known binding properties of the composite affinity reagents,
(ii) the binding pattern at each of the unique spatial addresses, and
(iii) a database of the plurality of different proteins.
15 . The method of claim 1 , wherein antibodies or fragments thereof in the plurality of antibodies or fragments thereof may be of the same type.
16 . The method of claim 1 , wherein antibodies or fragments thereof in the plurality of antibodies or fragments thereof may be of the differing types.
17 . The method of claim 1 , wherein the linker comprises a nanoparticle that is attached to the common identifiable tag.