Input: Expression data on tissue samples and a set of gene symbols that is known to be highly expressed in a specific cell type.
Output: Expression profile for each of the cell types in a tissue.
Step I: If W is known, proceed to step II, else estimate W using XS and equation 3.
Step II: Estimate S through quadratic programming.
where O is the gene expression profile on tissue samples, S is the expression profile for pure cell types, W is the weight matrix estimated using the marker genes, and t1 and t2 is the maximum and minimum measurable gene expression level. R package ‘quadprog’ is used to solve the quadratic programming problem.
Output: Expression profile for each of the cell types in a tissue.
Step I: If W is known, proceed to step II, else estimate W using XS and equation 3.
Step II: Estimate S through quadratic programming.
where O is the gene expression profile on tissue samples, S is the expression profile for pure cell types, W is the weight matrix estimated using the marker genes, and t1 and t2 is the maximum and minimum measurable gene expression level. R package ‘quadprog’ is used to solve the quadratic programming problem.
Free full text: Click here
