ERROR mu * T is too small

[dynamiskduo]

Hello, thank you for writing ProPIP, it sounds super cool! I am trying to use it on a fasta file with ~130 sequences and ~4000 characters as well as the correspondring tree.

When I start the run, it crashes after 2 %. The info file gives me this information

I20210302 11:53:56.959744 36695 main.cpp:837] [PIP model] Fixed PIP parameters to (lambda=15.8,mu=0.06,I=0.948)
I20210302 11:53:56.959820 36695 main.cpp:893] [Substitution model] Number of states: 5
I20210302 11:53:56.960119 36695 main.cpp:920] model=JC69+PIP(lambda=15.800000000000,mu=0.060000000000)
I20210302 11:53:56.960417 36695 main.cpp:968] [Alignment sequences] Starting MSA_t inference using Pro-PIP...
W20210302 11:53:56.962633 36695 progressivePIP.cpp:287] ERROR mu * T is too small: Success [0]
W20210302 11:53:56.962633 36695 progressivePIP.cpp:287] ERROR mu * T is too small: Success [0]
W20210302 11:53:56.962633 36695 progressivePIP.cpp:287] ERROR mu * T is too small: Success [0]
...
W20210302 11:53:56.962633 36695 progressivePIP.cpp:287] ERROR mu * T is too small: Success [0]

Do you know what causes this? Or do you have any ideas on how to work around it? I am not experienced in C++ so I am having a hard time trying to debug it, and any help is greatly appreciated.

Best, Amanda

[supermax1234]

Dear Amanda, this error was generated when the tool calculates the beta(v) probability for a given node v. The reason for the error is that the product of mu by the branch length is below a given threshold (1e-8). Most likely there is a branch length very close to zero.

Best, Max

[dynamiskduo]

Dear Max, thank you for your speedy reply :) Based on your advice, I pruned my tree while maximizing tree length and managed to avoid that error. Yay! However, the run still crashes after 15 %, but this time I do not get any warnings.

The screen output is:

[Computing the multi-sequence alignment]
Aligner optimised for:.................: ram
Stochastic backtracking active.........: no
[>>>>>>                                ]  15%Killed

and the log.INFO says:

Log file created at: 2021/03/06 15:18:54
Running on machine: g-03-c0181
Running duration (h:mm:ss): 0:00:00
Log line format: [IWEF]yyyymmdd hh:mm:ss.uuuuuu threadid file:line] msg
I20210306 15:18:54.041993 12366 main.cpp:265] alphabet:  DNA | gap-extention 1

Any advice here? Have a nice day, best, Amanda

[supermax1234]

Hi Amanda, this time I would try monitoring your RAM usage to check if you have enough. Best,

M

[dynamiskduo]

Hello again :) Hmmm, I am running on a cluster and allocate 120GB to analyse my pruned alignment w. 10 seqs. Is it expected to exceed that? Best, Amanda

[supermax1234]

Dear Amanda, try this option

alignment.version=cpu

in your configuration file.

Best,

Max

Il giorno lun 8 mar 2021 alle ore 15:01 dynamiskduo < notifications@github.com> ha scritto:

Hello again :) Hmmm, I am running on a cluster and allocate 120GB to analyse my pruned alignment w. 10 seqs. Is it expected to exceed that? Best, Amanda

— You are receiving this because you commented. Reply to this email directly, view it on GitHub https://github.com/acg-team/ProPIP/issues/1#issuecomment-792775308, or unsubscribe https://github.com/notifications/unsubscribe-auth/AGMTTV2Q64ZABOBDIO6U2SLTCTKE5ANCNFSM4YOYIPNQ .

[dynamiskduo]

Dear Max, thank you for your advice. Unfortunately, it still crashes after 15 %: [Computing the multi-sequence alignment] Aligner optimised for:.................: cpu Stochastic backtracking active.........: no [>>>>>> ] 15% And the error file says: *** Aborted at 1615283771 (unix time) try "date -d @1615283771" if you are using GNU date *** PC: @ 0x0 (unknown) *** SIGSEGV (@0x8) received by PID 23409 (TID 0x7ffff69c3600) from PID 8; stack trace: *** @ 0x7ffff47c5630 (unknown) @ 0x43da33 bpp::nodeCPU::DP3D_PIP_node() @ 0x454b1d bpp::CompositePIPnode::PIPnodeAlign() @ 0x44dcdb main @ 0x7ffff32e2555 __libc_start_main @ 0x45237f (unknown) /var/spool/torque/mom_priv/jobs/31087676.SC: line 28: 23409 Segmentation fault (core dumped) ProPIP params=indel_aware.txt I don't know if the two systems are just incompatible or something..

[supermax1234]

If you want, I can try to run the simulation on our cluster to see what happens.

Il giorno mar 9 mar 2021 alle ore 11:06 dynamiskduo < notifications@github.com> ha scritto:

Dear Max, thank you for your advice. Unfortunately, it still crashes after 15 %: [Computing the multi-sequence alignment] Aligner optimised for:.................: cpu Stochastic backtracking active.........: no [>>>>>> ] 15% And the error file says: Aborted at 1615283771 (unix time) try "date -d @1615283771" if you are using GNU date PC: @ 0x0 (unknown) SIGSEGV (@0x8) received by PID 23409 (TID 0x7ffff69c3600) from PID 8; stack trace: @ 0x7ffff47c5630 (unknown) @ 0x43da33 bpp::nodeCPU::DP3D_PIP_node() @ 0x454b1d bpp::CompositePIPnode::PIPnodeAlign() @ 0x44dcdb main @ 0x7ffff32e2555 __libc_start_main @ 0x45237f (unknown) /var/spool/torque/mom_priv/jobs/ 31087676.SC: line 28: 23409 Segmentation fault (core dumped) ProPIP params=indel_aware.txt I don't know if the two systems are just incompatible or something..

— You are receiving this because you commented. Reply to this email directly, view it on GitHub https://github.com/acg-team/ProPIP/issues/1#issuecomment-793656195, or unsubscribe https://github.com/notifications/unsubscribe-auth/AGMTTV5W77D2XNT3BCKF6TLTCXXIVANCNFSM4YOYIPNQ .