Today’s post, authored by a very smart guest cytogeneticist, nicely describes array-based comparative genomic hybridization, a very cool new DNA test that gives us a way to detect genetic abnormalities that are too small to be seen under the microscope. A student wanted to know more about why array CGH can only detect unbalanced rearrangements (like deletions) but not balanced rearrangements (like inversions).

Q. With regard to array CGH, I do not understand why balanced rearrangements could not be detected. Why can’t they make a probe for an inversion of a few genes or an insertion of a gene? I guess I do not see how these would be any different from making a probe for a deletion or duplication. I am sure I am missing something though.

A. Array-based CGH is a DNA based test that, in a much-simplified nutshell, looks at the quantity of DNA in a patient vs the quantity of DNA in a specimen derived from a pool of normal controls. Thousands of different probes from loci spanning the genome are present on a chip. If there is LESS DNA in the patient than the control for a particular probe, the a-CGH will show a DELETION of material from the patient for that particular locus; if there is MORE DNA in the patient than the control for a particular probe, the a-CGH will show a GAIN of material from the patient for that locus. In a balanced translocation, there is NO gain or loss of material, so the probes will show that the patient and the control have equal amounts of DNA in those translocated regions.

To detect a translocation, then, one would need to do a G-banded chromosomal analysis (i.e, look at the chromosomes under the microscope, the “old-fashioned” way). In that way, the material exchanged between the chromosomes involved in the translocation could be identified because they would LOOK different than their normal homologs — but because the translocation is balanced, there is NO gain or loss of DNA in this exchange, so array-CGH would not detect any genetic imbalance. In cancer cases, in which the genetic abnormalities involved in certain translocations have been well characterized (e.g., the 9;22 translocation in chronic myeloid leukemia involves breakage and rejoining of the ABL gene on chromosome 9q34 and the BCR gene on chromosome 22q11.2), FISH probes can be developed because we know the gene sequences of ABL and BCR. In contrast, however, for a constitutional balanced translocation that is passed on through a family or develops de novo in a patient, we don’t know what those genes are — so we don’t know the base-pair sequences that would enable us to develop a FISH probe. For these cases, then, we are limited to characterizing the abnormality as best we can by means of a G-banded chromosomal analysis.

At the risk of complicating this picture, I will add one further little scenario. Not all translocations that LOOK balanced in a G-banded chromosomal analysis really ARE balanced at the level of the DNA sequences. That is, in the process of formation of the translocation, sometimes very small amounts of DNA can be gained or duplicated at the breakpoints of the translocation. These amounts of DNA are way too small to be detected under the microscope in a G-banded chromosomal analysis (remember, the limit of detection for our eyes at the microscope is about 3 megabases of DNA (i.e. 3 million base pairs)). Gains or losses of DNA at the breakpoint of these apparently balanced translocations that are SMALLER than about 3 MB would NOT be detected in a G-banded chromosomal analysis. For this reason, if a patient is diagnosed with a de novo (not inherited from the parents) translocation that LOOKS balanced under the microscope (via a G-banded chromosomal analysis), we would still recommend that a-CGH be performed — because a-CGH would be able to detect any such small imbalances at the breakpoints of the translocation.

Image credit: DNA spiral by Charles Jencks, Kew Gardens, UK (by mira66,, under CC license).