Tsai asked in 科學生物學 · 1 decade ago


1.LOD是什麼?我有查到一點LOD score is the log of the ratio between the maximum likelihood that the data fit a null hypothesis and the maximum likelihood that the data fit a QTL hypothesis, given the observed phenotypes and marker genotypes.

QTL hypothesis是什麼呢? …文獻上說,GWAS基本上並不需要hypothesis,進行後續研究時才需要,那這裡的hypothesis是指什麼呢?

2.Minor allele frequency and effect sizes是什麼?和common and rare variation of SNPs對於GWAS結果呈現的影響?

3.關於 SNPs 的描述:我不太了解non-synonymous, non-conservative, synonymous, conservative


1 Answer

  • C C
    Lv 6
    1 decade ago
    Favorite Answer

    1. LOD score is a bit complicated to explain. For a simplified version, imagine this: you are trying to test for each marker in the genome its association with your phenotype of interest. Most of the markers in the genome will be unrelated to your phenotype, so they will just be background noises statistically. Then at markers that are actually associated with the phenotype, there will be a signal that rise above the noise. LOD score is a statistical value assigned to each marker to let you know how 'unlikely' it is that an association you see between this particular marker and your phenotype is due to random chance.

    2. in single nucleotide variations (SNPs), each marker has two alleles, say A and G. In a population of chromosomes, some chromosomes will have A, other chromosomes will have G. In the population the allele that are found less frequently is the minor allele, and its frequency in the population is the minor allele frequency.

    Effect size is just the magnitude of influence of a SNP to your phenotype. For example, if on average every allele of A increases your blood pressure by 3% standard deviation, then 3% is your effect size.

    In GWAS, it is much easier to find common variants associated with your phenotype, because you have more statistical power. Imagine that in 10,000 chromosomes (5,000 people, remember each person contributes 2 chromosomes), if you have an allele that is 30%, you would see the allele 3000 times. Statistically you would have much more confidence in drawing a conclusion based on something you see repeatedly over and over. If you only see the allele 1 time out of 10000, you don't know if the association you see is just by random chance, or is actually real.

    3. These definitions are with respect to a protein changes. A synonymous SNP means that even though there is a SNP, the actual amino acid is not affected, because of redundancy in amino acid codes. A non-synonymous SNP means that the amino acid is changed.

    2010-06-10 23:51:59 補充:

    (3. continued) Both conservative and non-conservative SNPs are non-synonymous SNPs. The difference is that conservative SNPs changed the amino acid,

    2010-06-10 23:52:13 補充:

    but would not have much influence on the protein structure because the one amino acid is replaced by another similar amino acid. Non-conservative SNPs would replace the amino acid with a very dissimilar amino acid, like a Ala -> Proline change or something like that.

    2010-06-10 23:52:44 補充:

    4. this figure is basically saying there are a lot of combination of SNPs with different frequency and different effects. Only those with STRONG effects are easy to pick up (upper left hand corner).

    2010-06-10 23:52:52 補充:

    But for common variants, because you see them so many times in a population, even if their effects are weaker you can still see them (lower right hand corner) in GWAS.

    2010-06-10 23:53:17 補充:

    The lower left hand corner variants will be very hard to find, because they are seen so few times in a population and also because they have very small effects.

    Source(s): i'm a graduate student in human genetics
Still have questions? Get your answers by asking now.