[Lewis Thomson (OP)]

Malcolm wrote in to the Naked Scientists to ask us this mutating mystery?

"How can we identify a disease? When looking at a sample of chromosomes, what are scientists looking for to spot a diseased gene?"

What do you think? Leave your answers in the comments below....

[Eternal Student]

Hi.

    You could be asking any of these things:

1.    What laboraratory techniques are used to sequence DNA and produce maps of a human genome.   What do the final results look like?   etc.

2.   What is the over-all principle for genetic testing used for diagnosis?

3.   How good is our theory matching DNA that a human has to characteristics they exhibit?

4.   Something else.

    I'm going to focus on  2 (Overall principles)  and my opinion of the limts of 3 (using pure theory to predict all characteristics from DNA).

- - - - - - -

     Genetic testing for diagnosing disease is mainly just the application of statistics to results obtained from a DNA tests that have been run on a larger group of people. 
     It isn't as if we have a perfect understanding of genetics or how genes are regulated and expressed.  We can't usually work from the map of a persons genes and just use pure theory to determine all their characteristics (like vulnerability to a disease).   For example, if a new gene mutation arises (like the Haemophilia gene that seemed to be a spontaneous mutation in the blood line of Queen Victoria of England), then we can not always predict ahead of time that this would cause disease.  We may only become aware of that AFTER the disease has been exhibited by the people and DNA testing was done which showed a strong correlation between having that mutated gene and having the disease.

This is usually the start of the process, the development of a genetic test:
   There is a large sample of people taken,   some with a known disease and some without.    Then their DNA is sampled and tested.   There are various features that can be identified from the collection of different testing processes they apply to the DNA.
     Then they apply some statistics to try and identify which features of that DNA seem to be associated with people who do and do not have that disease.   Where a statistical association is found between having the disease and some observable result from a test (like a band of colour appearing at a given place on a chromatogram),  then these will often be called "markers" for the disease.   With enough tests and work done, these "markers" can be specified right down to the detail of exactly which base pairs appear in the DNA and at which locations in the DNA - but often it's sufficient to leave it more vague than that.
     Sometimes a "marker" isn't just an indication that a person has (or does not have) one gene - sometimes to be vulnerable to a disease a person needs a whole collection of genes with a particular set of genotypes.   (To paraphrase this:  Sometimes they need a whole collection of genes to be set up a certain way).

After this initial work
    They now have some markers they can look for.  So they can sample the DNA of a person (who may not even be showing symptoms of the disease yet) and usefully tell them if they have the markers for the disease.   Depending on the statistical results they obtained when the markers were first identified PLUS the continuing development of the knowledge about the link between having those markers and developing the disease - the person can be advised appropriately.     
    For example,  it might be possible to tell them only a small amount of information   "We have found some markers for heart disease and based on our current data and understanding this puts you at double the risk of developing heart disease in later life.   However, it's still a small risk - you may never develop heart disease".   Sometimes, it's a lot of information or carries a much higher probability of diagnosis - for example the genetic causes of Cystic Fibrosis have been well studied and identified.   If some genetic markers for cystsic fibrosis are found then it is almost certain that the person will have symptoms from birth on-wards.

Best Wishes.

[evan_au]

It is sometimes hard to tell the difference between a disease and "normal variation". Most genetic mutations have no visible effect, and some make a small impact on normal variation. But if it doesn't affect your way of life, it isn't really a disease.

The simplest genetic diseases to diagnose are those cases where a disease is determined by 1 mutation at a particular point in the genome.
- However, for many conditions like hemophilia, or cystic fibrosis, there are many mutations in different parts of the genome that can produce the same general effect on the body (but perhaps with different severity)
- For the so-called "breast cancer genes" BRCA1 & BRCA2, there are specific mutations that increase your chance of cancer; but there are many other mutations which can occur in these same genes that might not impact your chance of breast cancer

Genes can be broadly classified as "dominant" (1 copy produces a change in the body) or "recessive" (1 "normal" gene will hide the effect).
- Dominant genes are easier to diagnose - a parent has some notable characteristic, and 50% of their children have the same notable  characteristic.
- Recessive genes are harder to diagnose - parents might show no symptoms of a particular characteristic, but 25% of their children have some notable characteristic. Depending on how common the recessive gene is in a particular population, the incidence of the disease may be anywhere from 1 in 100 to 1 in many millions.
- A special case that blurs these boundaries is where there are mutations on the X chromosome. Females have 2 X chromosomes, so a recessive mutation on one of them won't impact them. Males have only 1 X chromosome, so even a recessive mutation there will express itself. This effect is visible in cases of hemophilia, or some types of color blindness, where it affects males far more often than females.
https://en.wikipedia.org/wiki/X-linked_recessive_inheritance

The early diagnosis of genetic disease came from unusual symptoms carried in particular families.
- Sometimes that disease is more common in an entire population: eg having 1 copy of the sickle-cell gene provides some protection against malaria - so would you call that a disease at all? Unfortunately, having 2 copies of that gene causes severe problems with blood circulation.
- By testing many people who do or don't have the disease, researchers are able to home in on parts of the genome that are different between these two groups

With the reduction in the cost of whole-genome sequencing, the technique of "triple sequencing" is becoming more common
- When a baby is born with some unusual condition that can't be diagnosed, doctors will sequence the entire genome of the mother, father and baby
- The baby's genome is a combination of the mother and father's genome, so any place where the baby's genome differs from the parents is a possible "new" mutation that did not exist in either of the parents.
- By looking at the location of the mutation, researchers can sometimes tell the bodily function that might be affected by the mutation.
- Unfortunately, the function of many genes is still unknown, and mapping the location of regulatory areas is still a work in progress. Triple sequencing may give some clues about disease in newborns, and may assist in mapping these poorly understood regions of the genome.

[Eternal Student]

Hi.

   We seem to have good general agreement @evan_au.   There's just one point I would take a different line on.

However, for many conditions like hemophilia, or cystic fibrosis, there are many mutations in different parts of the genome that can produce the same general effect on the body (but perhaps with different severity)

   Haemophilia A is one of the easiest examples of disease with a clear genetic cause.  It is usually caused by a single faulty gene and is often used as a textbook example and studied in school.  That's why I used it as my example.
   Yes, there are complications (Haemophilia type B being an obvious example) but you're digging deep here and at some risk of confusing the situation.   Nothing is simple - but Haemophilia A is one of the most commonly studied examples of a disease with a simple genetic cause. 

   

Genes can be broadly classified as "dominant" (1 copy produces a change in the body) or "recessive" (1 "normal" gene will hide the effect).

    I think that's a considerably greater simplification of the situation by comparison.  It's also a quiet evening so I'm going to take a moment to moan about the gross simplifications that teachers make.

   Many phenotypes follow co-dominance of the individual alleles - none of the alleles are recessive or hidden, all of them will be expressed in the phenotype. Many more characteristics are the result of a complex interplay of several genes none of which show any clear dominant and recessive traits on their own.
    It's easier to teach about dominant and recessive genes which, I think, often leads students to believe that many characteristics should follow such a simple explanation.  In reality the textbooks and teachers have searched hard to find examples of characteristics that are so clearly caused by a Dominant and Recessive genes.  Most characteristics are not like this at all.   
    It's a bit like a Maths teacher teaching their young school students to measure the spread of data by calculating the range.   It's something you can teach easily but it's of incredibly little use in reality.   Measuring the spread by something like the standard deviation is far more useful but too difficult to teach and explain, so you just walk away from that.   Similarly Physics teachers will tell their students that gravity is a force,   Chemistry teachers say that mass is always conserved etc.  All science teachers simplify and sometimes mislead their students as a frequently un-intended consequence.   Suggesting that all genes can be broadly classified as dominant or recessive is very much in the same mode of teaching.  It's easy to teach but not particularly accurate.

NOTE:  I haven't studied Biology for years and am not claiming expertise here.  I'm also confident that @evan_au  isn't going to be worried about any challenge.  He/She is a moderator and already recognised as an expert in at least one area of science.

Best Wishes.