ERROR: Numerical underflow (lh-branch)

[Chih-Wei Huang]

Hello, 

I'm try to use IQTREE web server for evaluating the best substitution model and partition scheme.

It was returned with an error message.

This job STOPPED because there was a problem!

WARNING: Too many iterations in tqli
WARNING: Too many iterations in tqli
WARNING: Too many iterations in tqli
ERROR: Numerical underflow (lh-branch). Run again with the safe likelihood kernel via `-safe` option

I couldn't figure out what the problem is.

Thanks for your help.

Best regards,

Chih-Wei

The following is the running log 

IQ-TREE multicore version 1.6.11 for Linux 64-bit built Jun  6 2019
Developed by Bui Quang Minh, Nguyen Lam Tung, Olga Chernomor,
Heiko Schmidt, Dominik Schrempf, Michael Woodhams.

Host:    kronig (SSE4.2, 31 GB RAM)
Command: ../../iqtree -s 20190625_align_COIwithOG.fas -st DNA -sp COI_partition.txt -pre 20190625_align_COIwithOG.fas -m TESTONLYNEWMERGE
Seed:    351943 (Using SPRNG - Scalable Parallel Random Number Generator)
Time:    Tue Jun 25 15:39:17 2019
Kernel:  SSE2 - 1 threads (24 CPU cores detected)

HINT: Use -nt option to specify number of threads because your CPU has 24 cores!
HINT: -nt AUTO will automatically determine the best number of threads to use.

Reading partition model file COI_partition.txt ...
Reading alignment file 20190625_align_COIwithOG.fas ... Fasta format detected
Alignment most likely contains DNA/RNA sequences
Alignment has 150 sequences with 575 columns, 277 distinct patterns
253 parsimony-informative, 24 singleton sites, 298 constant sites
                                             Gap/Ambiguity  Composition  p-value
   1  B5                                             0.00%    passed     54.90%
   2  D3                                             0.00%    passed     54.90%
   3  VAL_00_Ea_Kar                                  0.00%    passed     36.91%
   4  VAR4                                           0.00%    passed     54.17%
   5  VAK_00_Xuyen_Moc                               0.00%    passed     62.54%
   .

   .

   .
 147  OG_BRAI02                                      0.00%    passed     83.99%
 148  OG_BRNS01                                      0.00%    failed      0.55%
 149  ZY4                                            0.00%    passed     13.50%
 150  OG_J                                           0.00%    failed      3.98%
****  TOTAL                                          0.00%  2 sequences failed composition chi2 test (p-value<5%; df=3)

Partition file is not in NEXUS format, assuming RAxML-style partition file...
Subset	Type	Seqs	Sites	Infor	Invar	Model	Name
1	DNA	150	192	191	0	GTR	part1
2	DNA	150	192	48	133	GTR	part2
3	DNA	150	191	14	165	GTR	part3
Degree of missing data: 0.000
Info: multi-threading strategy over partitions

NOTE: D3 is identical to B5 but kept for subsequent analysis
NOTE: YD1 is identical to R50 but kept for subsequent analysis
NOTE: YD7 is identical to YD2 but kept for subsequent analysis
NOTE: YD8 is identical to YD4 but kept for subsequent analysis
NOTE: XI3XI4XI6XI7XI8 is identical to XI0XI2 but kept for subsequent analysis
NOTE: R31 is identical to R30 but kept for subsequent analysis
NOTE: T71 is identical to T70 but kept for subsequent analysis
NOTE: X61X62X63X64 is identical to X50-1-2-3-4-5-7 but kept for subsequent analysis
NOTE: T30 is identical to T21 but kept for subsequent analysis
NOTE: XA3567 is identical to XA1 but kept for subsequent analysis
NOTE: VAY0 is identical to S51 but kept for subsequent analysis
NOTE: VAO1 is identical to VAG_00_Chu_Prong but kept for subsequent analysis
NOTE: XX6 is identical to XX0123457 but kept for subsequent analysis
NOTE: T61 is identical to T60 but kept for subsequent analysis
NOTE: X24 is identical to X21 but kept for subsequent analysis
NOTE: XV0 is identical to X23 but kept for subsequent analysis
NOTE: XH0XH2XH6789 is identical to XG1XG2XG8 but kept for subsequent analysis
NOTE: XG4XG5 is identical to X22 but kept for subsequent analysis
NOTE: XH4 is identical to XG7XG9 but kept for subsequent analysis
NOTE: XV6 is identical to XG6 but kept for subsequent analysis
NOTE: ZF9 is identical to ZE8 but kept for subsequent analysis
NOTE: VAU1 is identical to T41 but kept for subsequent analysis
NOTE: 16 identical sequences (see below) will be ignored for subsequent analysis
NOTE: YD3 (identical to R50) is ignored but added at the end
NOTE: YE4 (identical to R50) is ignored but added at the end
NOTE: YE1 (identical to YD2) is ignored but added at the end
NOTE: YE3 (identical to YD2) is ignored but added at the end
NOTE: T31 (identical to T21) is ignored but added at the end
NOTE: XW1XW3 (identical to XA1) is ignored but added at the end
NOTE: VBA4 (identical to S51) is ignored but added at the end
NOTE: T81 (identical to T60) is ignored but added at the end
NOTE: X25 (identical to X21) is ignored but added at the end
NOTE: XG0XG3 (identical to X21) is ignored but added at the end
NOTE: XH1XH5 (identical to X21) is ignored but added at the end
NOTE: XV3 (identical to X21) is ignored but added at the end
NOTE: XV7 (identical to X21) is ignored but added at the end
NOTE: XV2XV9 (identical to X23) is ignored but added at the end
NOTE: XV5 (identical to XG1XG2XG8) is ignored but added at the end
NOTE: XV1 (identical to XG7XG9) is ignored but added at the end
Concatenated alignment was printed to 20190625_align_COIwithOG.fas.uniqueseq.phy

For your convenience alignment with unique sequences printed to 20190625_align_COIwithOG.fas.uniqueseq.phy


Create initial parsimony tree by phylogenetic likelihood library (PLL)... 0.018 seconds
NOTE: ModelFinder requires 12 MB RAM!
Testing GTR+F+G on supermatrix...
GTR+F+G4 / LnL: -9241.935 / df: 274 / AIC: 19031.870 / AICc: 19534.204 / BIC: 20224.968
Selecting individual models for 3 charsets using BIC...
 No. Model        Score       Charset
   1 TIM2+F+ASC+G4 14954.630   part1
WARNING: Too many iterations in tqli
WARNING: Too many iterations in tqli
WARNING: Too many iterations in tqli
ERROR: Numerical underflow (lh-branch). Run again with the safe likelihood kernel via `-safe` option

[Chih-Wei Huang]

I've try again. No error message.

Thanks!

Chih-Wei Huang於 2019年6月25日星期二 UTC+8下午9時40分49秒寫道：

[Minh Bui]

Hi Chih-Wei,

Just as a head up: I saw that part1 has no invariant sites. This might or might not be problematic, but you should check the data carefully.

Cheers,

Minh

-- 
You received this message because you are subscribed to the Google Groups "IQ-TREE" group.
To unsubscribe from this group and stop receiving emails from it, send an email to iqtree+un...@googlegroups.com.
To post to this group, send email to iqt...@googlegroups.com.
Visit this group at https://groups.google.com/group/iqtree.
To view this discussion on the web visit https://groups.google.com/d/msgid/iqtree/0c974298-dfe2-4174-9b8a-7a4fdab60855%40googlegroups.com.
For more options, visit https://groups.google.com/d/optout.