Karyotypes? FISH? Arrays? What is this, the 20th century???

One of the biggest subspecialties in clinical genetics is cytogenetics.

It is the field within pathology that deals with cellular anatomy as it relates to chromosomes.

For the longest time, humanity’s best guess at figuring out genetic disease was either to look at the physical attributes of a person (phenotype) or look at the chromosomal super-structures within cells using a microscope.

Cytogenetics today is mostly focused on pre-natal/post-natal clinical testing in fetuses and infants.

These are some of the first tests that are performed when there's a suspected genetic disease.

You might be thinking, "We have sequencing now, so what's the point of looking at chromosomes?"

Because structural variants like inversions, translocations and deletions/duplications are a big deal, but doing anything using sequencing in cytogenetics is a very recent development and this is underscored by the fact that up until about 2005, archaic methods like metaphase karyotyping were still the gold standard cytogenetic technique.

But, in the 21st century, we can do better, faster, cheaper with far higher resolution!

For comparison, the traditional methods for looking at chromosomes are:

Karyotyping: Either stain based banding or more recent molecular banding based techniques with a resolution of about 5 megabases (mb).

FISH: Fluorescent in-situ hybridization which is where fluorescently labeled probes are used to detect specific chromosomal abnormalities or paint whole chromosomes.

aCGH: Array comparative genomic hybridization uses probe array technology to detect genomic anomalies with 50-100kb resolution.

"But what about non-invasive prenatal testing, isn't that sequencing???"

Yes, but it’s not diagnostic, and the resolution of NIPT is on it’s best day 5mb. It also cannot reliably detect inversions or translocations - because, short-reads.

"Ok, then what about whole genome sequencing?"

Same problem as above but with slightly better resolution. If you're doing WGS anyway, short-read based cytogenetic predictions can RULE IN a diagnosis but they can't be used to rule out one because short-reads miss things.

Enter high resolution genomic mapping.

The leader in this space is Bionano Genomics and they use two enzymes to fluorescently tag DNA.

The DNA is then stretched out, linearized, and imaged in megabase sized chunks.

The little florescent tags can then be used to reconstruct, or map, a genome with a resolution of 3-10kb, picking up everything that's missed by competing methods for a couple hundred dollars.

Nabsys now does something similar with a pore based detection scheme.

Or, if you have a penchant for doing sequencing, you can spend a bit more to use Hi-C based techniques (Arima, Phase Genomics, etc) to make similar reconstructions.

But the biggest barrier here for broad clinical use will be convincing payers to reimburse for them!