-
Notifications
You must be signed in to change notification settings - Fork 0
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
table of dn/ds values seem odd #24
Comments
I think this is fixed now. The issue was, I think, when two sequences were equal, therefore with a dN and dS of zeros. I have solved this by skipping those pairs. Still if you want to inspect this better, here's the pairwise values:
|
Close it if you agree this is fixed. |
ping Can I close this? |
I had a look at this and the values people seem to quote are 10-20x higher. See:
There is possibly some variation by the counting algorithm, and your dN and dS values will be 100x lower because you are per site whereas they are counting events in the sequence (103 aa). It seems you have dN = 0.0134 x 103 = 1.4 which is close enough to their 0.7. However, dS = 2.76 x 103 = 280 which is 5x their value. I will think some more about this ... |
I just realised that based on the IDs you seem to have done a NP v NP (protein) comparison. Is that right? Can you please change this to a gene coding region (DNA) based calculation. Should be straightforward? |
The calculations were done with the DNA coding sequences. The IDs were of the proteins because the alignment is done in protein space and then the CDS were mapped to it (otherwise the alignment would space 3 nucleotide gaps instead of a single codon gap). |
I sent the manuscript on to Cathal Seoighe to get his feedback |
I have updated the table S8 legend based on Cathal's feedback. Two additional items to fix in scripts:
|
Why? Is this due to propagation of rounding errors? Maybe I'm doing it wrong but I think this is more. If mean is a sum of a list and then division by an exact number, then the result should be rounded to the number of significant figures of the sum of the values. The number of significant figures of that sum is dependent on the least significant digit of the series of numbers being summed (and not their number of significant digits). For example, for H3 dN, the number with least precision is significant to the 1/1000 digit (all the dN numbers are like that). The sum of the dN until that digit is 0.3687. This number has 4 significant digits so after division should be rounded to 4 significant digits and so to 0.004097 And for dS (which has bigger numbers), the sum is 227.8282 which has 7 significant digits so after division with the same number of significant digits is 2.531424.
Fixed. |
Values on the table of dn/ds values seem odd. I'm to geta list of dn, ds, and omega values for each each H2A pair to seee what's going on. Andrew to review it.
The text was updated successfully, but these errors were encountered: