Naked Science Forum
Life Sciences => Plant Sciences, Zoology & Evolution => Topic started by: lyner on 23/07/2009 15:39:37
-
I have looked all over the place about this but there seems to be a hole in the popular literature (or my sample of it).
My question is that I can read about and now understand the principle behind the statistics of inheritance. I can also understand how they can map the genome and the mechanics of replication. What I can't get my head round is where they get the connections between all the alleles and the actual, physical, strands of DNA that carry the particular length of gene.
If there were two individuals identical, but for one characteristic, then the difference between the genetic information could be connected with certainty. I can even see how the statistics of huge samples can connect all sorts of cause / effect information in medical surveys; again, the numbers are comprehensible. But approaching the genetic code this way seems far too complicated.
A five line explanation would be appreciated.
-
Hi SC
These days it's much easier than it used to be. Historically the only way to track down a gene was to start with the protein that it encoded, work out the protein sequence, then, using the genetic code (what triplets encode which amino acids) work out the possible gene sequences that could encode that protein (there will be more than one because of the degeneracy of the DNA code), and then probe the human genome looking for the location of those gene sequences.
So if you wanted to know what chromosome it was on, for instance, you could assemble a karyotype (you line up all the chromosomes next to each other so you can identify them) having used a colour-tagged gene probe to find where the sequence you're interested in sits on the chromosome.
These days it's a lot easier to locate genes linked to diseases or family traits. Researchers use a trick called genome-wide analysis. Here thousands of individuals with a certain condition are compared with thousands of normal "control" individuals. What researchers are looking for are features called SNPs (pronounced "snips") - single nucleotide polymorphisms - that crop up more often in people with the condition than without it. SNPs are effectively genetic flags - they are short pieces of sequence mapped to known locations on the genome. Because DNA sequences that are close together tend to be inherited together (ie are resistant to crossing over) if a certain SNP is much more common in people with a given condition then it's very likely that sitting near that SNP in the genome is a gene or piece of code that has some effect on the condition.
Because we know where the SNPs all are, and we know the human genome sequence, you can now map the condition to a genomic region and then look at the sequence to spot candidate genes linked to your disease.
Chris
-
That's brilliant, Chris. Thanks.
How easy is it to link a protein with a characteristic? Would it be as straightforward as identifying the protein in a skin pigment? But what about something like six fingers? Which protein?
For the second method, it just looks like an immense amount of processing of vast amounts of data. I guess, from what you are saying, that it is do-able! Ain't computers great?
-
To track down the gene that's responsible for a trait isn't too tricky. If a gene is playing a role in making a cell behave in a certain way then that gene must be switched on in the target cell. Therefore, capturing all of the gene products (a species of genetic material called mRNA - messenger RNA) from a cell will yield, amongst others, that gene product. Then it's a case of weeding out which gene (from the thousands) is the one you want.
One way to do this is to build a gene library. You insert all of the gene products into bacterial plasmids and then grow them in the bacteria to bulk up the numbers. You then divide the plasmids into pools of say 100. You then assay (test) for the gene of interest amongst these pools to find which pool has bacteria in it that make the gene product. You then start to subdivide that pool further until you find the bacterial colony making the gene product you want.
Chris
-
How would that be achieved for a Number of Fingers gene?
I can see how it could work where a 'chemical effect' can be identified but not when there isn't anything as 'obvious'.
-
Scientists can create "knockout mice," which have been mentioned many times on The Naked Scientists. Basically, they manipulate the genome of a single mouse egg before it divides and take a gene out. Whatever that mouse can't do, the removed gene is responsible for.