pathtalk.org is a weblog about pathology.

Testing-testing: CGH me baby!!

Testing-testing: CGH me baby: I want to know how I measure up to the reference.

This is a very superficial post touching on an increasingly utilized testing methodology.

Comparative genomic hybridization (CGH) (also has been called “competitive”) is a method for scanning the entire genome for variations in DNA copy number which is not currently common in the clinical practice of medicine, but can be a useful tool in the molecular study of tumors. CGH can also be attempted in the search for chromosomal aberrations in patients with congenital anomalies of unknown etiology when other standard methods such as cytogenetics have not yielded a satisfactory result.

Tumor or germ-line DNA can be studied using this method. Genomic DNA is isolated from both the patient (and tumor) and a reference population. Both the test and the reference metaphase genomic DNA is labeled with fluorescent probes with different fluorescent characteristics. This mix of reference DNA, test DNA, and hybridized fluorescent probes are studied using fluorescent microscopes and equipment with the ability to measure the intensity of fluorescent signals. If the test DNA has signal which is much greater than the reference DNA (increased copy number) this will be detected by a change in signal compared to the expected value. Conversely, if the test DNA has decreased copy number of a particular gene compared to the reference population, then the fluorescent signal will be altered and detected. CGH will detect extra chromosomes and missing or altered chromosomes.

Array CGH is a CGH method in which the genomic DNA is hybridized to a microarray. This has improved the detection limit on this test; however the single copy number aberration must occur over a region of about 10 Million base pairs. Methods utilizing a focused microarray for hybridization can improve this detection limit further.

Some argue this may completely replace the discipline of cytogenetics because the methodology is steadily improving and its flexible. You can look at multiple focused aspects of the genome or the entire genome. The current resolution is supposed to be about 10,000 base pairs, but currently the signal noise associated with this resolution confounds its utility. However, cytogenetics will not be replaced just yet, as this method does not confer information about mosaicism and would be expensive to implement currently.

References
Nature Genetics 37, S11 - S17 (2005)
Oncol Rep 2007; 18:917-26.

Gretchen - another interesting post. A couple of things to consider - the future of array based CGH is oligonucleotide arrays rather than the BAC arrays you describe. These have the POTENTIAL to have a resolution of hundreds of bases rather than hundreds of kilobases. You are correct that arrays will never replace convential cytogenetics, however. As good as they will be (and detecting mosaicism is certainly possible with a good high density oligo array), the one thing an array can never detect is a balanced translocation.

BTW, shouldn’t this be CGH my baby (with unexplained developmental delay)?

ahh totally!!

CGH my baby!! nice!!

this is how i have seen it used the most currently

i didnt know mosaicism is possible!! that is cool!!
i was asaking around and people didnt seem to think so –probably becuase it is still a send out for us

so the oligoarrays would improve detection? cool i need to read about this

another thing i am really unfamiliar with about arrays in general and this probably applys to CGH intensely (and am interested) in is data processing and interpretation

if i had to imagine how it is done and reading the little bits of info you can find in articles –there is computer analysis for over and underexpression– then correlations are determined based on these patterns-again probably computer generated

so if you have a panel of speciffic genes or chromosome regions which are over/under to a varying degree from say 100 patients how do you interpret the data of a few genes here and there different than those in the pilot studies

much less how is this data stored and tracked over a period of years– example repeat CGH in a leukemia patient to monitor changes in the tumor or for clonal evolution like some people use cytogenetics

Processing and interpretation is the hardest part of an array based assay (and often the part that is relegated to a ‘black box’) Differences in background subtraction and data normalization can have an enormous impact on the ultimate interpretation of an array. Conceptually, two color BAC arrays are easier to understand (and interpret) than the one color high density oligo arrays (which are much noisier on the level of individual probes, and being one color, have no internal control for direct comparison or dye switching experiments).

Of course this is only the math stuff - you have to remember that a larger than you would expect chunk of the human genome has copy number heterogeneity across individuals or populations (I think the number is 7% - or is that the proportion of developmental delay caused by subtelomeric rearrangements? Anyway, it’s in that range 7-10 ish). So even if you see a difference in copy number in a baby or a tumor, it may be a normal variant rather than a disease causing ‘mutation’. There is a fine study on this topic in an article in Nature in 2006 (or 2007) which specifically looked at copy number variance using oligo and BAC arrays.