Internal ERROR when running aBSREL

[Joye-YF-Zhou]

Hello HyPhy team,

I just ran an aBSREL analysis on a given dataset and at the "Improving parameter estimates of the adaptive rate class model " step, it says: Internal error

Internal error 

ERROR: [_LikelihoodFunction::LocateTheBump (index 304) current 
   value   -126826.3114144693 (parameter =   0.9993584395344067), best value   
   -126826.3114144693 (parameter =   0.9993584395344067)); delta = 
  1.455191522836685e-11, tolerance = 9.999999999999 
  Parameter name 
 GZjCEMgC.tree_0.Eupolybothrus_cavernicolus_TRINITY_DN19382_c0_g2_i1_len935_.omega2

 Function call stack
1 :  [namespace = vPhaUHwH] Optimize(mles, ^lf_id);

-------
 2 :  
 absrel.full_model.fit=estimators.FitExistingLF(absrel.likelihood_function_id,absrel.model_object_map);
-------

Is there any reason why this would happen?

information on how aBSREL is ran:

• HyPhy version 2.5.48
• downloaded through conda
• command: absrel --alignment MUSCLE.txt --tree MUSCLE.contree
• This issue is only seen on this given alignment. The other datasets have been able to run aBSREL successfully

I have attached the aBSREL run log here. The alignment could be emailed to upon request.

Thank you for looking over this error absrel_run.log.zip

[spond]

Dear @Joye-YF-Zhou,

Please rerun the analysis specifying ENV="TOLERATE_NUMERICAL_ERRORS=1;" on the command line. Also see https://github.com/veg/hyphy/issues/1630

I made some fixes in subsequent versions (2.5.53 is current), which may reduce the incidence of the warning/error.

Best, Sergei

[Joye-YF-Zhou]

Hello Dr. Pond,

Thank you for the reply! I would rerun my analysis with the specifications.

Best regards, Joye Zhou

[github-actions[bot]]

Stale issue message