IQ-TREE crash

[Cuong Dang]

Hi,

I got a crashing message when running iqtree-sse3 with a big protein alignment of 4899 taxa.

I paste on the end some last lines of the log.

Best regards,

Cuong Dang.

Create initial parsimony tree by phylogenetic likelihood library (PLL)... 178.252 seconds

Reading model parameters from file JTT.paml

NOTE: 4720.754 MB RAM is required!

Estimate model parameters (epsilon = 0.001)

iqtree: iqtree-1.1.3-Source/phylotreesse.cpp:1611: double PhyloTree::computeLikelihoodDervEigenTipSSE(PhyloNeighbor*, PhyloNode*, double&, double&) [with VectorClass = Vec2d; int VCSIZE = 2; int nstates = 20]: Assertion `isnormal(tree_lh)' failed.

*** IQ-TREE CRASHES WITH SIGNAL ABORTED

*** For bug report please send developers:

***    Log file: HIV_train.phylip.log

***    Alignment files (if possible)

Aborted (core dumped)

[Bui Quang Minh]

Dear Cuong,

thanks for reporting this bug. Essentially it is a numerical underflow (even with proper numerical scaling) due to such a huge alignment. RAxML published this technical issue, if you want to know details:

http://www.biomedcentral.com/1471-2105/12/470

I also ran RAxML on your alignment and it gave the same behaviour, i.e. sometimes tree log-likelihood becomes -inf.

So this issue will not likely be resolved soon, unfortunately.

Minh

--
You received this message because you are subscribed to the Google Groups "IQTREE" group.
To unsubscribe from this group and stop receiving emails from it, send an email to iqtree+un...@googlegroups.com.
Visit this group at http://groups.google.com/group/iqtree.
For more options, visit https://groups.google.com/d/optout.

--

Bui Quang Minh

Center for Integrative Bioinformatics Vienna (CIBIV)

Max F. Perutz Laboratories

Campus Vienna Biocenter 5, VBC5, Ebene 1

A-1030 Vienna, Austria

Phone: ++43 1 4277 24026

Fax: ++43 1 4277 24098

Email: minh.bui (AT) univie.ac.at

[chank...@gmail.com]

I got the same problem when running iqtree on a large dataset of 11000 SNP loci. It crashed during the model selection part. Below is the last few lines of the output:

26 K3P+I 2937269.815 190 5874919.629 5874919.672 5877265.369
27 K3P+G4 2944672.181 190 5889724.362 5889724.405 5892070.102
28 K3P+I+G4 2934922.544 191 5870227.089 5870227.132 5872585.175
29 K3Pu 2920397.859 192 5841179.718 5841179.762 5843550.150
30 K3Pu+I 2893714.106 193 5787814.211 5787814.256 5790196.989

*** IQ-TREE CRASHES WITH SIGNAL ABORTED

*** For bug report please send to developers:
*** Log file: /scratch/chan/iqtree/md5mc48p.10.phy.log

*** Alignment files (if possible)

Has there been a solution for this problem yet?

[Karen Meusemann]

Hi all,

I have the same problem doing partitoned BS analyses (15 taxa, 260000 sites, aa dataset, 234 partitions...
anyway, I definitely need to analyse this - so shall I resample that long until it does not crash?
(I run 10 runs a 10 bootstrap replicates; one (until now) was successful)

what would yu suggest: other partition scheme?
other commands?
I would like to do the normal BS though...

Best Karen

WARNING: Numerical underflow caused by alignment sites 2 29 48 66 126 152 155 199 248 285 319 346 380 392 416 439
WARNING: Numerical underflow caused by alignment sites 2 152 199 248 346 392 416
WARNING: Numerical underflow caused by alignment sites 2 199 416
iqtree: .../iq/iqtree-1.3.10-Source/phylotreesse.cpp:2108: double PhyloTree::computeMixrateLikelihoodBranchEigen(PhyloNeighbor*, PhyloNode*): Assertion `!isnan(tree_lh) && !isinf(tree_lh)' failed.

*** IQ-TREE CRASHES WITH SIGNAL ABORTED

*** For bug report please send to developers:

***    Log file: rcl_iq_bs30.log

***    Alignment files (if possible)

/cm/local/apps/slurm/var/spool/job10275850/slurm_script: line 12: 98172 Aborted                 iqtree -omp 20 -st AA -s ...supermatrix_aa_alltaxa.phy -spp ...coll_partitions_rcl.nex -pre rcl_iq_bs00 -bo 10 -numstop 300 -sprrad 10

If I shall send you the alignment etc please let me know! (quite urgent unfortunately so for any hints running thorough BS I am very gratefull)

[Bui Quang Minh]

Dear Karen, 

yes please send me via personal email the data, command, and log file. I will have a closer look.

Thanks, Minh

To post to this group, send email to iqt...@googlegroups.com.

Visit this group at http://groups.google.com/group/iqtree.
For more options, visit https://groups.google.com/d/optout.

--

Bui Quang Minh

Center for Integrative Bioinformatics Vienna (CIBIV)

Campus Vienna Biocenter 5, VBC5, Ebene 1

A-1030 Vienna, Austria

Phone: ++43 1 4277 74326

Email: minh.bui (AT) univie.ac.at

[sanchezarr...@gmail.com]

Hello,

I just got the same error:

*** IQ-TREE CRASHES WITH SIGNAL ABORTED
*** For bug report please send to developers:

*** Log file: iq_.log

*** Alignment files (if possible)

Abort trap: 6

But my alignment is only 38 taxa with only 4173 nucleotide positions, not so big at all, I don't know why it is crashing.

To call the software I used: iqtree-omp -s papara_alignment_basalGB.fasta -m TIM3+R5 -alrt 1000 -bb 1000 -bo 100 -pre iq_ -wsr -wbt -nt 6

I attach the log file in case you might need it.


Thank you very much for your help,


Alicia.

[Bui Quang Minh]

Hi Alicia,

please do not perform both ultrafast bootstrap (-bb) and standard bootstrap (-bo) within one run. It crashed at the final step because of this. I will add a check in the future release to disable this.

Minh

> <iq_.log>