Substitution rate error

[spotto]

Github issue: There is something wrong with the scale of duotang’s “Molecular clock estimates”, which suggests ~1500 substitutions per genome per day. That’s way too high. It should be closer to 0.06 substitutions per genome per day (about one change every two weeks). Did this accidentally get multiplied by the genome size? <@U02NQPKG3UJ> <@U03VC9B7PC3>

Slack Message

[ArtPoon]

Ok I'll check it out

[ArtPoon]

I suspect the problem is that when I replaced the RTT plot with an interactive JS widget, the code for calculating clock rates did not get updated. It is probably pulling raw DOM coordinates instead of the divergence (y-axis) values.

[ArtPoon]

Ok this was a pretty easy fix. @spotto 's intuition was correct - for some reason rescaling the rate estimates by genome length became superfluous, so I am removing this from the dev branch:

diff --git a/scripts/fit-rtt.R b/scripts/fit-rtt.R
index a5e539f5..ec012382 100644
--- a/scripts/fit-rtt.R
+++ b/scripts/fit-rtt.R
@@ -80,9 +80,9 @@ get.ci <- function(fits) {
   ci <- lapply(fits, confint.rlm)
   est <- data.frame(
     n = sapply(fits, function(f) nrow(f$x)),                            
-    est = 29903*sapply(fits, function(f) f$coef[2]),
-    lower.95 = 29903*sapply(ci, function(f) f[2,1]),
-    upper.95 = 29903*sapply(ci, function(f) f[2,2])
+    est = sapply(fits, function(f) f$coef[2]),
+    lower.95 = sapply(ci, function(f) f[2,1]),
+    upper.95 = sapply(ci, function(f) f[2,2])
   )
   est$Lineage <- row.names(est)
   est

[bfjia]

Thank you Art