Calculating Ti/Tv ratio

[Heisman]

Hey all,

Quick question. When, for example, it's stated that for human exomes a Ti/Tv ratio of around 3.0 is expected, does that refer to the number of transitions divided by the number of transversions, or the number of transitions multiplied by 2 divided by the number of transversions (ie, absolute counts or rates)? I've been googling around and I think it's rates but I want to be absolutely sure that the commonly reported expected ratios refer to rates as well. Thanks for any help.

[SES]

It sounds like you are confusing the Ti/Tv ratio with the ratio of the Ti/Tv rates. The first calculation you mention would be like something you would do to get the ratio of the rates. The Ti/Tv ratio is simply the ratio of the number of transitions to the number of transversions for a pair of sequences. Be sure to pay attention to the model of evolution used in the studies you are referring to, otherwise you may not get the same results.

[Heisman]

It sounds like you are confusing the Ti/Tv ratio with the ratio of the Ti/Tv rates. The first calculation you mention would be like something you would do to get the ratio of the rates. The Ti/Tv ratio is simply the ratio of the number of transitions to the number of transversions for a pair of sequences. Be sure to pay attention to the model of evolution used in the studies you are referring to, otherwise you may not get the same results.

Ok, so for example, if I have 100 SNPs, 75 of them are transitions, and 25 are transversions, then the Ti/Tv ratio would be 75:25 = 3.0? Based on the last paragraph here http://www.broadinstitute.org/gsa/wi...php/QC_Methods , it looks like that calculation (75/25) would suffice and give me an expected value for an exome.

But then, reading the second reply here http://www.biostars.org/post/show/47...ral-rule/#4758 indicates I should be multiplying by a factor of 2, if I understand correctly. Hence, it would be 75*2/25 = 6.0

I'm still not sure if I should throw in that factor of 2 when comparing to established whole genome or whole exome values.

[SES]

I believe the discussion you are referring to in the latter link is about the observed difference in Ti/Tv rates for the whole genome vs. those rates when only looking at exome data. The first link also has a discussion of this point in that final paragraph. I don't think you should be multiplying by any factor. There is a worked example of this calculation on the snpEff homepage that may clarify this point.

[Heisman]

I believe the discussion you are referring to in the latter link is about the observed difference in Ti/Tv rates for the whole genome vs. those rates when only looking at exome data. The first link also has a discussion of this point in that final paragraph. I don't think you should be multiplying by any factor. There is a worked example of this calculation on the snpEff homepage that may clarify this point.

Ah, cool. No multiplication factor then. Thanks!