System and method for classifying a body tissue sample

A program storage device readable by machine, tangibly embodying a program of instructions executable by the machine to perform method steps for classifying a test tissue sample into a class from among K classes, the test tissue sample having a gene expression vector {tilde over (X)}=({tilde over (x)}1, . . . {tilde over (x)}N) where {tilde over (x)}j is an expression level of a gene j in the test tissue sample, comprising:

• a computer-readable medium comprising:
• (a) for k=1 to K
• (i) for each of Mk tissue samples in the class k,
• (I) for each of the N genes, calculating a centralized matrix Xk− Mk, where Xk=(Xijk), a gene expression matrix in which Xijk is an expression level of the jth gene in the ith tissue sample, where j=1 to N and i=1 to Mk;
• and where
• M
• _
• k
• =
• 1
• /
• M
• k
• *
• ∑
• i
• =
• 1
• M
• ⁢
• ⁢
• X
• i
• k
• ,
• ⁢
• and
• ⁢
• ⁢
• X
• i
• k
• =
• (
• x
• i
• ⁢
• ⁢
• l
• k
• x
• i
• ⁢
• ⁢
• 2
• k
• ⋮
• x
• iN
• k
• )
• ;
• (ii) a first instruction for calculating a covariance matrix Ck of Xk− Mk;
• (iii) a second instruction for calculating one or more eigenvectors of the matrix Ck;
• (iv) a third instruction for calculating a metric μk indicative of an extent of dissimilarity between the vector {tilde over (X)} and the matrix Xk, the metric μk being calculated using an algebraic expression involving one or more eigenvectors of the matrix Cx, and
• (b) classifying the test tissue sample into a class {tilde over (k)}, where {tilde over (k)} is a class for which the metric μk is minimum among the k classes.

A computer program product comprising a computer readable medium comprising computer readable program code embodied therein for classifying a test tissue sample into a class from among K classes, the computer program product comprising:

• (a) computer readable program code, for k=1 to K
• (i) for each of Mk tissue samples in the class k,
• (I) for each of the N genes, to calculate a centralized matrix Xk− Mk, where Xk=(Xijk), a gene expression matrix in which Xijk is an expression level of the jth gene in the ith tissue sample, where j=1 to N and i=1 to Mk;
• and where
• M
• _
• k
• =
• 1
• /
• M
• k
• *
• ∑
• i
• =
• 1
• M
• ⁢
• X
• i
• k
• ,
• and
• ⁢
• ⁢
• X
• i
• k
• =
• (
• x
• il
• k
• x
• i
• ⁢
• ⁢
• 2
• k
• ⋮
• x
• iN
• k
• )
• ;
• (ii) a first instruction to calculate a covariance matrix Ck of Xk− Mk;
• (iii) a second instruction to calculate one or more eigenvectors of the matrix Ck;
• (iv) a third instruction to calculate a metric μk indicative of an extent of dissimilarity between the vector {tilde over (X)} and the matrix Xk, the metric μk being calculated using an algebraic expression involving one or more eigenvectors of the matrix Cx, and
• (b) a further instruction to classify the test tissue sample into a class {tilde over (k)}, where {tilde over (k)} is a class for which the metric μk is minimum among the k classes.