• Question: How can geneticists tell what part of the DNA sequence correlates to different attributes?

    Asked by to Reka, Judith, Gemma, Aoife on 3 Jan 2019. This question was also asked by .
    • Photo: Gemma Chandratillake

      Gemma Chandratillake answered on 3 Jan 2019: last edited 3 Jan 2019 6:17 pm

      We used methods like “linkage analysis” and “genome-wide association studies” to correlate a characteristic to a particular location in the genome. Basically, we look for genetic landmarks that coincide with the characteristic in a family or a population.

    • Photo: Reka Nagy

      Reka Nagy answered on 4 Jan 2019:

      A bit more detail about the ‘genome-wide association studies’, or GWAS, that Gemma mentioned:
      To do this, you need a number of people (the more, the better, some data such as the UK Biobank Resource has over 500,000 participants!), and from these people you need two things:

      THEIR DNA
      * This can be either in the form of genome sequencing, where you read every letter in their 3-billion letter-long DNA – this is more expensive so it is not usually done at the hundreds of thousands scale –
      * Or it can be in the form of genotypes – where you only look at a subset of all DNA letters (variants), usually on the scale of hundreds of thousands to millions. Which letters to look at is generally determined by several things – for example, you may be interested in those letters that, when mutated, are known to be involved in causing disease, or those letters for which two different variants exist within the population.

      THEIR TRAIT(S)
      You also need to measure one or more traits in these people. Such traits can be, for example,
      – ‘external’, such as height, weight or blood pressure
      – ‘internal’ such as blood sugar levels
      – behavioural such as the number of times you exercise a week, or the amount of TV you watch
      – disease-related, such as whether you have, for example, diabetes or cancer

      Once you have these two things, you can look at each DNA letter you’ve measured in turn, and look at how which version (allele) of the DNA letter each person has correlates with an observed trait.

      For example, if people who are taller than average all have an ‘A’ allele , and people who are shorter than average all have a ‘T’ allele, while those in the middle have one of each, you might conclude that this variant may have an impact on height! Computers can do the grunt work of doing this comparison across millions of letters in millions of people, in many different traits, and the hard work is actually in trying to interpret the results into something meaningful and informative!

Comments