Effect Of Mutation On Codon

Question

[Gvj]

Hi All, Do you know any general scoring method/matrix which can give a score to a mutation based on its probability to change amino acids and the nature of changes (eg: acid to basic or vice versa is more lethal than basic to basic changes ) it makes ? This is on a genome scale and truly based on codon rather than the genomic location.

Eg: A to G change (Transition) will affect all codons containing A, but degeneracy of the codon will maintain same amino acids (eg: UUA and UUG produces Leucine). So considering all the changes one mutation could make, I want to give a score to each changes and compare between different mutations. Is there any scoring matrix/method exists for this ?

Thanks in advance,

[Hayssam]

Do you want to do that on a specific organism or on the genetic code?

Answer 1

[Jordan]

I'm not quite clear about the question. But there are general scoring matrices such as BLOSUM and PAM which give a score for amino acid substitutions. You have different scores for different levels of evolution as well (i.e., based on evolutionary distance, like PAM1 for closely related sequences and PAM250 for distantly related sequences).

But if you just want to predict if the sequence is synonymous/non-synonymous mutation we can use, markov chains. I 'think' Forward algorithm might work, because you are trying predict the observations (in this case synonymous(S) and non-synonymous(N)).

Answer 2

[biorepine]

Try SIFT

[Gvj]

Thanks, but as far as I understood, SIFT is to find the effect of mutation on a "particular protein" by aligning similar protein to it. What I need is the effect of one type of mutation on genomic scale. In other words, a score for a mutation based on the number of amino acid changes it could make (based on codons) and the impact of each amino acid changes based on its physicochemical properties. Once I have a score for each type of mutation, then I could say one kind of mutation (eg A to T) is more severe than another kind (eg :A to G). I hope I am clear

[Alex Paciorkowski]

I don't understand what you mean by "the effect of one type of mutation on genomic scale." Nonsynonymous nucleotide changes are constrained by the codon and the gene in which they are found. Any attempt to conclude whether one mutation is more severe than another will be based upon the biology of the specific protein. Therefore, SIFT, polyphen, etc will be the algorithms that will most helpfully determine this -- barring getting into specific protein-based experiments.

Answer 3

[Vivek]

If the mutation is non synonymous coding, you can find the mutation effect scores across different prediction tools from dbNSFP for different alternate alleles and protein changes.

https://sites.google.com/site/jpopgen/dbNSFP

Answer 4

[AsoInfo]

You can try polyphen, mutation taster, provean for the amino acid substitution changes.

Answer 5

[Gvj]

I was looking for a matrix like Grantham matrix . Your replies helped me to find this paper and some other related articles.

Thanks for help