Parameters to root and generate time caliberated phylogenies for dataset with weak temporal signals

[Rohit-Satyam]

Hi Treetime Developers

I am creating Dengue 1,2 and 3 phylogenies. When running the TreeTime, i get strong temporal signal for Dengue 1 and Dengue 3 but not for Dengue 2.

treetime clock --tree iqtree_denv2.treefile --dates denv2_treetime.csv --aln denv2_cialign_cleaned.fasta \
--reroot least-squares --outdir 01_denv2_treetime --name-column name --date-column date --clock-filter 0 \
--allow-negative-rate

0.00	-TreeAnc: set-up

Attempting to parse dates...
	Using column 'name' as name. This needs match the taxon names in the tree!!
	Using column 'date' as date.

12.87	TreeTime.reroot: with method or node: least-squares

12.90	TreeTime.reroot: rerooting will ignore covariance and shared ancestry.

 Root-Tip-Regression:
 --rate:	7.643e-05
 --r^2:  	0.01

The R^2 value indicates the fraction of variation inroot-to-tip
distance explained by the sampling times.Higher values corresponds
more clock-like behavior (max 1.0).

The rate is the slope of the best fit of the date tothe root-to-tip
distance and provides an estimate ofthe substitution rate. The rate
needs to be positive!Negative rates suggest an inappropriate root.


The estimated rate and tree correspond to a root date:

--- root-date:	 -946.30


--- new tree written to 
	01_denv2_treetime/rerooted.newick

--- wrote dates and root-to-tip distances to 
	01_denv2_treetime/rtt.csv

--- root-to-tip plot saved to  
	01_denv2_treetime/root_to_tip_regression.pdf

With this knowledge of presence of weak temporal signal, how should I proceed? I also wish to help confidence values but I can't specify --covariation parameter as per documentation.

treetime --tree 01_denv2_treetime/rerooted.newick --dates denv2_treetime.csv --aln denv2_cialign_cleaned.fasta --name-column name --date-column date --keep-root --time-marginal only-final --confidence --outdir 02_denv2_timetree --covariation --stochastic-resolve

With the command above I get the following plot:

While trying to run without --covariation, treetime says:

Outside of covariation aware mode TreeTime cannot estimate confidence
intervals without specified standard deviation of the clock
rate.Please specify '--clock-std-dev' or rerun with '--covariation'.
Will proceed without confidence estimation

I am not sure what value should I pass to --clock-std-dev!!

Edit1 I also tried other methods other than least-squares but the plots didn't change much.

[Rohit-Satyam]

Tried with custom clock-rate of --clock-filter 0 --clock-rate 6e-4 and I get the following plot. So does it make sense?

Another one with reported DENV2 nucleotide substitution rated reported in a paper here: --clock-filter 0 --clock-std-dev 7e-06 --clock-rate 6.93e-4