Help with divergence time estimation

[Samuel Bernardes]

Hi everyone,

I'm new to MCMCtree, and I want to date a phylogeny containing 148 terminals and obtained from a genomic alignment 1,773,009 bp long. I've followed tutorial 2 (MCMCTree tutorials) and the book chapter, but my results don't make much sense.

My output file looks incomplete, but the run is stated as finalised. The out.txt contains only the alignment and patterns and its last lines are:

Homogeneity statistic: X2 = 0.01317 G = 0.01316

Average                        0.25997 0.21506 0.29566 0.22931
(Ambiguity characters are used to calculate freqs.)

# constant sites:  35835 (2.02%)

However, there's another output file written by Slurm. This seems more similar to what I'm supposed to expect (e.g. it contains a line 'Time used'). The complete file is attached, but its last lines are:

Time used: 2:05:59
es (79.5%),  1:09Collecting fpatt[] & pose[], 1266924 patterns at 1420000 / 1773009 sites
(80.1%),  1:09Collecting fpatt[] & pose[], 1266924 patterns at 1430000 / 1773009 sites
(...)
(99.3%),  1:09Collecting fpatt[] & pose[], 1266924 patterns at 1770000 / 1773009 sites
(99.8%),  1:09Collecting fpatt[] & pose[], 1266924 patterns at 1773009 / 1773009 sites
(100.0%),  1:09Counting frequencies..
1266924 patterns, messy

The .ctl file as well as the batch file are also attached below. Any help is welcome.

A few points:

- I have only one calibration point near the root and the time unit is supposed to be in million years

- The analysis never lasted for more than 2 hours, regardless of the time limit I set in the batch file.

Thank you in advance,

Sam

[Sishuo Wang]

Hi Sam,

Thanks for your question! I wondered if you mind uploading (or sending me via email) all input files and outputs so that i can have a closer look.

best,

sishuo

[Samuel Bernardes]

Hi Sishuo,

I've made a Google Drive folder (https://drive.google.com/drive/folders/163hpMWcj_yl582uLbRc6TDKpcmchC3YH?usp=drive_link) with all the outputs and inputs. I removed some parts of the output and changed the names of the terminals for confidentiality and to reduce their size since some files repeat the alignment all over. In case you need them, I can share the inputs privately with you.

Thanks for the help.

Sam

[Sishuo Wang]

Solved. thx for asking. further Qs, pls feel free to discuss on this forum! many people are here to help.

Hi Sishuo,

 

Thanks a million for your help. I figured out the problem. It was a problem with the cluster. I don't know what exactly was happening, but there was something wrong with the tmp files and the analysis was finishing halfway through.

 

Again, I appreciated the attention spent on my problem.

 

All the best,
Sam