At long last -- I can finally talk about the huge genomics project I've been working on!
Xenografts are human tumors that are grown in mice. They are important models for learning about the development of cancer and for testing in vivo drug efficacy and stuff. An increasingly common application is to take a piece of tumor from someone, grow the pieces up in a number of mice and figure out which chemotherapy drug (or combination of drugs) works best on this particular tumor. Powerful tools, right?
Well, we wondered 1) how different xenografted tumors were from liquid growing cell lines and 2) how stable xenografts are through passages; in other words, how stable are xenografts if grow one in a mouse and then take some of that and move it to another mouse. By passaging xenografts we can get more material while limiting the suffering of individual mice.
The easiest way to get a feel for how these cell lines differ? A shit ton of genomics via microarray. My job was analysis and quality control of the arrays. Did I mention there were a lot? Yeah....for just the first 49 cell lines completed (there were going to be 50, but we dropped HeLa on ethical grounds) there are 823 arrays, each representing about 14,000 genes. This project is on-going. Eventually there will be thousands of arrays, with a series for every cell line studied by the National Cancer Institute. This paper is an introduction to the project and the an overview of the first cohort.
What we've found out so far: some cells are super stable. Some cells differentiate like crazy and will be very different after a few passages. PCA analysis can be used to determine what cell lines are permeated by mouse tissue and that it may be possible to track sensitivity to some drugs across cohorts this large when the pathways of sensitivity/resistance are well understood.
You can check out this paper at BMC Genomics here.
If you are really interested, you can download the RAW array data for every sample set here.
I wonder if there is any material left over from this study to do proteomics on...
No comments:
Post a Comment