US 12,460,260 B2
Methods utilizing single cell genetic data for cell population analysis and applications thereof
Aaron M. Newman, San Mateo, CA (US); Gunsagar S. Gulati, Roxbury, MA (US); Michael F. Clarke, Menlo Park, CA (US); and Shaheen Shabbir Sikandar, Mountain View, CA (US)
Assigned to The Board of Trustees of Leland Stanford Junior University, Stanford, CA (US)
Filed by The Board of Trustees of the Leland Stanford Junior University, Stanford, CA (US)
Filed on May 26, 2020, as Appl. No. 16/883,915.
Claims priority of provisional application 62/852,231, filed on May 23, 2019.
Prior Publication US 2020/0370112 A1, Nov. 26, 2020
Int. Cl. C12Q 1/6869 (2018.01); C12N 5/00 (2006.01); C12N 15/113 (2010.01); C12Q 1/6883 (2018.01)
CPC C12Q 1/6869 (2013.01) [C12N 5/0068 (2013.01); C12N 15/113 (2013.01); C12Q 1/6883 (2013.01); C12N 2506/30 (2013.01)] 18 Claims
OG exemplary drawing
 
1. A method to infer a differentiation status and identify a differentiation-status biomarker from single-cell RNA sequencing of a tissue comprising a mixed population of cells in which the biomarker is prognostic of disease outcome, comprising:
obtaining single cell RNA transcriptome sequencing results of tissues comprising a mixed population of cells, wherein the sequencing was performed on each single cell individually and each cell's sequencing result includes the cell's gene expression;
generating a matrix of cell×cell's gene expression for the population of cells from the sequencing results;
determining a gene count signature for the population of cells, wherein the gene count signature is a surrogate for whole transcriptome gene counts and defined as a geometric mean of the expression level of each gene of a set of five or more genes that are most correlated with gene counts within the population of cells;
computing a gene count signature metric for each cell;
determining the differentiation status of each cell of the population of cells relative to each other based on ranking each cell's smoothed gene count signature metric, wherein cells having a smoothed gene count signature metric indicative of lower gene counts are more differentiated than cells having a smoothed gene count signature metric indicative of higher gene counts;
wherein each cell's smoothed gene count signature metric comprises denoising the gene count signature metric by:
identifying transcriptionally similar cells within the population of cells utilizing a neighboring technique and the matrix of cell×cell's gene expression; and
smoothing the gene count signature metric by:
applying a regression technique to determine coefficients that best fit the gene count signature to the neighboring technique; and
simulating a diffusion process to iteratively adjust the gene count signature for a number of iterations or until convergence;
identifying a differentiation-status biomarker that is expressed in a subset of cells of the population of cells, wherein each cell of the subset of cells is determined to have the same differentiation status by their smoothed gene count signature metric rank, wherein the differentiation-status biomarker is uniquely expressed within the subset of cells as compared to the whole population of cells establishing that the differentiation-status biomarker is a gene product indicative of the differentiation status of the subset of cells, wherein the subset of cells has a putative effect on disease prognosis based on its differentiation status;
determining the biomarker is prognostic of disease by modulating gene expression of the biomarker in an experimental model that is representative of the disease;
performing a prognostic assay on a patient tissue sample comprising a mixture of cells having varying degrees of differentiation, wherein the prognostic assay comprises one of the following:
detecting RNA expression of the biomarker within the patient tissue sample by nucleic acid hybridization; or
detecting protein expression of the biomarker within the patient tissue sample by immunodetection.