Compositions and methods for defining cells

1. A computer-implemented method of assaying a cell comprising:

providing said computer with a reference database, wherein the reference database comprises cell datasets produced from characteristic data including global profile information for at least one-hundred fifty known cell lines, tissues, or primary cells and at least one dataset produced for each of human embryonic stem cells, human induced pluripotent stem cells and human somatic multipotent cells;

receiving a test dataset in said computer, wherein the test dataset comprises data including global profile information for a cell line, tissue, or primary cell of previously unknown pluripotency;

merging the test dataset into the reference database producing an updated reference database; and

performing in said computer unsupervised classification of the updated reference database thereby producing a computed label classification of the test dataset, wherein the computed label classification includes an indication of whether or not the cell line, tissue, or primary cell in the test dataset is pluripotent or contains pluripotent cells.

2. The method of claim 1 , further comprising the step of performing automatic model selection producing a best fitting classification model, wherein the best fitting classification model is a model that an algorithm identifies as most stable to random perturbations.

3. The method of claim 2 , further comprising identifying in which cluster or clusters the unknown cell is grouped.

4. The method of claim 3 , further comprising outputting a computed definition of the unknown cell wherein the computed definition is referenced to the cluster in which the unknown cell resides.

5. The method of claim 4 , wherein the best fitting classification model can cluster the individual datasets such that each dataset within a cluster is indistinguishable from each other dataset within the cluster.

6. The method of claim 4 , wherein the best fitting classification model can cluster the individual datasets such that each dataset within a cluster is similar to each other individual dataset in the cluster.

7. The method of claim 6 , further comprising identifying computationally derived class labels based only on biological characteristics.

8. The method of claim 7 , further comprising identifying differences in at least one dataset for at least one label between at least two samples in at least two clusters.

9. The method of claim 8 , further comprising filtering within a cluster for samples having a similar label profile.

10. The method of claim 9 , further comprising defining differentially regulated protein-protein networks.

11. The method of claim 10 , further comprising using the protein-protein networks to define a class membership, manipulate class membership, or define biological function of an unknown cell.

12. The method of claim 4 , wherein the best fitting classification model can cluster the individual datasets such that each dataset within a cluster is different from each other individual dataset.

13. The method of claim 1 , wherein performing unsupervised classification of the updated reference database comprises clustering RNA, DNA and/or protein profiles.

14. The method of claim 1 , wherein the global profile information is obtained from microarray analysis of cellular RNA.

15. The method of claim 1 , wherein performing unsupervised machine classification comprises a bootstrapping sparse non-negative matrix factorization.

16. The method of claim 1 , wherein the reference database comprises transcriptional profiles for stem cells.

17. The method of claim 1 , further comprising classifying cells based on computationally derived protein-protein network.

18. The method of claim 1 , wherein the global profile is a transcriptional profile.

19. The method of claim 1 , wherein the reference database comprises known class labels.

20. The method of claim 1 , further comprising the step of outputting results from the unsupervised classification.

21. The method of claim 1 , wherein the reference database is provided on a storage medium.

22. The method of claim 1 , wherein receiving the test dataset comprises receiving input from a computer system.

23. The method of claim 1 , wherein receiving the test dataset comprises receiving input from an array analysis system.

24. The method of claim 1 , wherein receiving the test dataset comprises receiving input via a computer network.

25. The method of claim 1 , wherein the data in the reference database is associated with one or more labeled associated biological classes of the cells.

26. A computer program product comprising a non-transitory machine readable medium on which is provided program instructions, the program instructions comprising:

code for receiving a test dataset, wherein the test dataset comprises data including global profile information for a cell line, tissue, or primary cell of previously unknown pluripotency;

code for merging the test dataset into the reference database producing an updated reference database, wherein the reference database comprises cell datasets produced from characteristic data including global profile information for at least one-hundred fifty known cell lines, tissues, or primary cells and at least one dataset produced for each of human embryonic stem cells, human induced pluripotent stem cells and human somatic multipotent cells;

code for performing unsupervised classification of the updated reference database producing a computed label classification of the test dataset, wherein the computed label classification includes an indication of whether or not the cell line, tissue, or primary cell in the test dataset is pluripotent or contains pluripotent cells.

27. A computer-implemented method comprising:

providing said computer with a reference database, wherein the reference database comprises cell datasets produced from characteristic data including global profile information for at least one-hundred fifty known cell lines, tissues, or primary cells and at least one dataset produced for each of human embryonic stem cells, human induced pluripotent stem cells and human somatic multipotent cells;

receiving a test dataset in said computer, wherein the test dataset comprises data including global profile information for a cell line, tissue, or primary cell of previously unknown pluripotency;

merging the test dataset into the reference database producing an updated reference database; and

performing in said computer unsupervised classification of the updated reference database thereby producing a computed label classification of the test dataset, wherein the computed label classification includes an indication of whether or not the cell line, tissue, or primary cell in the test dataset has a biological characteristic of interest.