Minh 2020: IQ-TREE 2, example.phy tutorial data

The session

This is the session as the scientist sees it, in order. Decision cards show the answer that the script gave and where that answer comes from. Each step shows the program, its version, the input file hash and the outputs. Click a line to see more.

- Setup values:
Alignment:
- Alignment method (MAFFT) (algorithm): auto
- Trim the alignment (trim): false
Tree:
- Substitution model (model): MFP
- Bootstrap type (bootstrap_type): ufboot
- Number of bootstrap replicates (replicates): 1000
- Outgroup (outgroup): 
- Random seed (seed): 1
Ask the scientist: Largest gap fraction of a kept column (max_gap_fraction).
Fixed by the lab: Random seed (seed) = 1.
The harness fills these values into your tool calls automatically. If you give a different value, the harness logs a deviation and shows it to the scientist.
- Mode is now watch: you run the steps.

report={work}/infer_tree-1/run.iqtree
seed=1
bootstrap_type=ufboot
bootstrap_replicates=1000
treefile={work}/infer_tree-1/run.treefile
consensus_tree={work}/infer_tree-1/run.contree
n_sequences=17
n_sites=1998
model=TIM2+F+I+G4
best_model_bic=TIM2+F+I+G4
log_likelihood=-21152.5377
tree_length=4.2173
constant_sites=686
site_patterns=1152

(LngfishAu:0.1709068277,(LngfishSA:0.1883311672,LngfishAf:0.1648032873)100:0.1073256416,(Frog:0.2561720790,((((Turtle:0.2214955116,(Crocodile:0.3056436573,Bird:0.2309752831)97:0.0651735079)70:0.0364941210,Sphenodon:0.3446303789)51:0.0204677635,Lizard:0.3859794405)99:0.0739928283,(((Human:0.1850870817,(Seal:0.0943875559,(Cow:0.0822572767,Whale:0.1012437124)99:0.0404174962)71:0.0252594799)93:0.0340514112,(Mouse:0.0583901122,Rat:0.0905005039)100:0.1217432345)100:0.0607172631,(Platypus:0.1919645817,Opossum:0.1509515712)96:0.0373002662)100:0.1490260063)100:0.1275594673)100:0.0940391076);

IQ-TREE 3.1.4 built Sep 10 2026

Input file name: {data}/minh2020-iqtree2/example.phy
Type of analysis: ModelFinder + tree reconstruction + ultrafast bootstrap (1000 replicates)
Random seed number: 1

REFERENCES
----------

To cite IQ-TREE 3 please use:

Thomas K.F. Wong, Nhan Ly-Trong, Huaiyan Ren, Piyumal Demotte, Hector Banos,
Andrew J. Roger, Edward Susko, Chris Bielow, Nicola De Maio, Nick Goldman,
Matthew W. Hahn, Mario dos Reis, Le Sy Vinh, Gavin Huttley, Robert Lanfear, Bui Quang Minh (2026)
IQ-TREE 3: Phylogenomic Inference Software using Complex Evolutionary Models.
Molecular Biology and Evolution, msag117.
https://doi.org/10.1093/molbev/msag117

Since you used ModelFinder please also cite: 

Subha Kalyaanamoorthy, Bui Quang Minh, Thomas KF Wong, Arndt von Haeseler,
and Lars S Jermiin (2017) ModelFinder: Fast model selection for
accurate phylogenetic estimates. Nature Methods, 14:587–589.
https://doi.org/10.1038/nmeth.4285

Since you used ultrafast bootstrap (UFBoot) please also cite: 

Diep Thi Hoang, Olga Chernomor, Arndt von Haeseler, Bui Quang Minh,
and Le Sy Vinh (2018) UFBoot2: Improving the ultrafast bootstrap
approximation. Mol. Biol. Evol., 35:518–522.
https://doi.org/10.1093/molbev/msx281

SEQUENCE ALIGNMENT
------------------

NOTE: Alignment sequence type is auto-detected. If in doubt, specify it via -st option.
Input data: 17 sequences with 1998 nucleotide sites
Number of constant sites: 686 (= 34.3343% of all sites)
Number of invariant (constant or ambiguous constant) sites: 686 (= 34.3343% of all sites)
Number of parsimony informative sites: 1009
Number of distinct site patterns: 1152

ModelFinder
-----------

Best-fit model according to BIC: TIM2+F+I+G4

List of models sorted by BIC scores: 

Model                  LogL         AIC      w-AIC        AICc     w-AICc         BIC      w-BIC
TIM2+F+I+G4      -21152.739   42383.479 +   0.0741   42385.072 +   0.0828   42601.875 +    0.974
GTR+F+I+G4       -21148.957   42379.913 +    0.441   42381.674 +    0.453   42609.509 -   0.0214
TIM2+F+R3        -21150.544   42383.088 +   0.0901   42384.849 +   0.0926   42612.684 -  0.00438
GTR+F+R3         -21147.067   42380.134 +    0.395   42382.071 +    0.372   42620.930 - 7.09e-05
TPM2u+F+R3       -21283.140   42646.279 - 6.36e-59   42647.955 - 6.82e-59   42870.275 - 5.08e-59
TPM2u+F+I+G4     -21291.534   42659.067 - 1.06e-61   42660.580 - 1.24e-61   42871.863 -  2.3e-59
TVM+F+R3         -21279.783   42643.567 - 2.47e-58   42645.414 - 2.43e-58   42878.762 -  7.3e-61
TVM+F+I+G4       -21287.937   42655.873 - 5.25e-61   42657.549 - 5.63e-61   42879.870 -  4.2e-61
TIM2+I+G4        -21308.990   42689.980 - 2.06e-68   42691.338 - 2.59e-68   42891.576 -  1.2e-63
TIM2e+I+G4       -21309.009   42690.018 - 2.02e-68   42691.376 - 2.54e-68   42891.614 - 1.18e-63
TIM2e+I+G4       -21309.009   42690.018 - 2.02e-68   42691.376 - 2.54e-68   42891.614 - 1.18e-63
TIM2+R3          -21303.599   42683.199 - 6.12e-67   42684.712 - 7.12e-67   42895.995 - 
... (974 more characters in the session record)

- 3 numbers in the answer match no logged tool result: 95. Run the tool that measures each number, or remove the number, or say how you derived it from logged results.
Fix these problems. Then give the answer again.

The final answer

The model wrote this answer. The harness shows it to the scientist after the review.

Settings used, from the decision record: Substitution model: MFP · Bootstrap type: ufboot · Number of bootstrap replicates: 1000 · Random seed: 1.

Checks

Review findings

The review recorded 8 findings. A rule finding comes from a fixed check in the harness. A referee finding comes from a second model that reads the record. The harness shows the findings to the scientist with the final answer. The record does not mark a finding as fixed. Thus a finding from an early review round can apply to a draft that the model corrected later.

Numbers in the answer

The last claim check read 47 numbers in the answer. 44 numbers match a logged result. 3 numbers have no source in the record.

Deviations

The model did not try to change a choice of the scientist.

Failed tool calls

No tool call failed.

Data integrity

Each data file has the same SHA-256 hash now as at the time of the step that read it. Where the download script (fetch.sh) gives a hash, the file also has that hash. The run did not change the data.

A SHA-256 hash is a fingerprint of the file contents. If one byte of the file changes, the hash changes. The table shows the first 12 characters.

How to repeat it

Get the data. The script downloads the files and checks their SHA-256 hashes where it lists them.

CUVETTE_DATA={data} bash bench/papers/minh2020-iqtree2/fetch.sh

Run the same case with Cuvette. The script gives the same answers from bench/papers/minh2020-iqtree2/bench.yaml.

cuvette bench papers --papers minh2020-iqtree2 --models claude:claude-opus-5-5

Repeat each step by hand in the program. For each step, the harness records a manual route: the menu path or the code that gives the same result. This list does not include comparison runs.

1. infer_tree (step n1)

   Run: iqtree3 -s <alignment> -m <model> -B <replicates> -seed <seed> -T 1

   • -s

     {data}/minh2020-iqtree2/example.phy

   • -m = MFP
   • -B = 1000
   • -seed = 1
   • Warning: If you keep the default random, you get a different result.

   The manual route that the harness recorded

   /bin/sh {other volume}/tools/overnight/claude-final/catalog/phylo/scripts/iqtree.sh /opt/homebrew/bin/iqtree3 infer --alignment {data}/minh2020-iqtree2/example.phy --model MFP --bootstrap-type ufboot --replicates 1000 --seed 1

   The manual route gives the same numbers. An automatic test in Cuvette checks this.

Figure

Run facts

The session

This is the session as the scientist sees it, in order. Decision cards show the answer that the script gave and where that answer comes from. Each step shows the program, its version, the input file hash and the outputs. Click a line to see more.

- Setup values:
Alignment:
- Alignment method (MAFFT) (algorithm): auto
- Trim the alignment (trim): false
Tree:
- Substitution model (model): MFP
- Bootstrap type (bootstrap_type): ufboot
- Number of bootstrap replicates (replicates): 1000
- Outgroup (outgroup): 
- Random seed (seed): 1
Ask the scientist: Largest gap fraction of a kept column (max_gap_fraction).
Fixed by the lab: Random seed (seed) = 1.
The harness fills these values into your tool calls automatically. If you give a different value, the harness logs a deviation and shows it to the scientist.
- Mode is now watch: you run the steps.

report={work}/infer_tree-1/run.iqtree
seed=1
bootstrap_type=ufboot
bootstrap_replicates=1000
treefile={work}/infer_tree-1/run.treefile
consensus_tree={work}/infer_tree-1/run.contree
n_sequences=17
n_sites=1998
model=TIM2+F+I+G4
best_model_bic=TIM2+F+I+G4
log_likelihood=-21152.5377
tree_length=4.2173
constant_sites=686
site_patterns=1152

(LngfishAu:0.1709068277,(LngfishSA:0.1883311672,LngfishAf:0.1648032873)100:0.1073256416,(Frog:0.2561720790,((((Turtle:0.2214955116,(Crocodile:0.3056436573,Bird:0.2309752831)97:0.0651735079)70:0.0364941210,Sphenodon:0.3446303789)51:0.0204677635,Lizard:0.3859794405)99:0.0739928283,(((Human:0.1850870817,(Seal:0.0943875559,(Cow:0.0822572767,Whale:0.1012437124)99:0.0404174962)71:0.0252594799)93:0.0340514112,(Mouse:0.0583901122,Rat:0.0905005039)100:0.1217432345)100:0.0607172631,(Platypus:0.1919645817,Opossum:0.1509515712)96:0.0373002662)100:0.1490260063)100:0.1275594673)100:0.0940391076);

- The answer names standard for "Bootstrap type", but the decision record says ufboot. Report the value that was used.
- 2 numbers in the answer match no logged tool result: 95. Run the tool that measures each number, or remove the number, or say how you derived it from logged results.
Fix these problems. Then give the answer again.

The final answer

The model wrote this answer. The harness shows it to the scientist after the review.

Settings used, from the decision record: Substitution model: MFP · Bootstrap type: ufboot · Number of bootstrap replicates: 1000 · Random seed: 1.

Checks

Review findings

The review recorded 8 findings. A rule finding comes from a fixed check in the harness. A referee finding comes from a second model that reads the record. The harness shows the findings to the scientist with the final answer. The record does not mark a finding as fixed. Thus a finding from an early review round can apply to a draft that the model corrected later.

Numbers in the answer

The last claim check read 36 numbers in the answer. 32 numbers match a logged result. 4 numbers have no source in the record.

Deviations

The model did not try to change a choice of the scientist.

Failed tool calls

No tool call failed.

Data integrity

Each data file has the same SHA-256 hash now as at the time of the step that read it. Where the download script (fetch.sh) gives a hash, the file also has that hash. The run did not change the data.

A SHA-256 hash is a fingerprint of the file contents. If one byte of the file changes, the hash changes. The table shows the first 12 characters.

How to repeat it

Get the data. The script downloads the files and checks their SHA-256 hashes where it lists them.

CUVETTE_DATA={data} bash bench/papers/minh2020-iqtree2/fetch.sh

Run the same case with Cuvette. The script gives the same answers from bench/papers/minh2020-iqtree2/bench.yaml.

cuvette bench papers --papers minh2020-iqtree2 --models claude:claude-sonnet-5-5

Repeat each step by hand in the program. For each step, the harness records a manual route: the menu path or the code that gives the same result. This list does not include comparison runs.

1. infer_tree (step n1)

   Run: iqtree3 -s <alignment> -m <model> -B <replicates> -seed <seed> -T 1

   • -s

     {data}/minh2020-iqtree2/example.phy

   • -m = MFP
   • -B = 1000
   • -seed = 1
   • Warning: If you keep the default random, you get a different result.

   The manual route that the harness recorded

   /bin/sh {other volume}/tools/overnight/claude-final/catalog/phylo/scripts/iqtree.sh /opt/homebrew/bin/iqtree3 infer --alignment {data}/minh2020-iqtree2/example.phy --model MFP --bootstrap-type ufboot --replicates 1000 --seed 1

   The manual route gives the same numbers. An automatic test in Cuvette checks this.

Figure

Run facts

The session

This is the session as the scientist sees it, in order. Decision cards show the answer that the script gave and where that answer comes from. Each step shows the program, its version, the input file hash and the outputs. Click a line to see more.

- Setup values:
Alignment:
- Alignment method (MAFFT) (algorithm): auto
- Trim the alignment (trim): false
Tree:
- Substitution model (model): MFP
- Bootstrap type (bootstrap_type): ufboot
- Number of bootstrap replicates (replicates): 1000
- Outgroup (outgroup): 
- Random seed (seed): 1
Ask the scientist: Largest gap fraction of a kept column (max_gap_fraction).
Fixed by the lab: Random seed (seed) = 1.
The harness fills these values into your tool calls automatically. If you give a different value, the harness logs a deviation and shows it to the scientist.
- Mode is now watch: you run the steps.

report={work}/infer_tree-1/run.iqtree
seed=1
bootstrap_type=ufboot
bootstrap_replicates=1000
treefile={work}/infer_tree-1/run.treefile
consensus_tree={work}/infer_tree-1/run.contree
n_sequences=17
n_sites=1998
model=TIM2+F+I+G4
best_model_bic=TIM2+F+I+G4
log_likelihood=-21152.5377
tree_length=4.2173
constant_sites=686
site_patterns=1152

(LngfishAu:0.1709765394,(LngfishSA:0.1884042664,LngfishAf:0.1648765322)100:0.1073673156,(Frog:0.2562938256,((((Turtle:0.2215450915,(Crocodile:0.3057507446,Bird:0.2310620493)97:0.0651606330)70:0.0364894400,Sphenodon:0.3447780527)51:0.0204536475,Lizard:0.3861439885)99:0.0739888076,(((Human:0.1851778377,(Seal:0.0944230016,(Cow:0.0823003161,Whale:0.1012908438)99:0.0404371901)71:0.0252744924)93:0.0340669342,(Mouse:0.0584259465,Rat:0.0905415196)100:0.1218030843)100:0.0607194581,(Platypus:0.1920645910,Opossum:0.1510258428)96:0.0373287871)100:0.1490890925)100:0.1275997766)100:0.0940810002);

import re
path = "{data}/minh2020-iqtree2/example.phy"
with open(path) as f:
    lines = [l.rstrip("\n") for l in f if l.strip()]
header = lines[0].split()
n, L = int(header[0]), int(header[1])
seqs = []
for l in lines[1:]:
    parts = l.split(None, 1)
    seqs.append(parts[1] if len(parts) > 1 else "")
# interleaved format: join all residue chunks per sequence
if len(seqs) > n:
    seqs2 = [""] * n
    for i, l in enumerate(lines[1:]):
        parts = l.split(None, 1)
        if i < n:
            seqs2[i] = parts[1].replace(" ", "") if len(parts) > 1 else ""
        else:
            seqs2[i % n] += l.replace(" ", "")
    seqs = seqs2
seqs = [s.replace(" ", "") for s in seqs]
total = sum(len(s) for s in seqs)
gaps = sum(s.count("-") + s.count("?") for s in seqs)
print("sequences", n, "declared sites", L)
print("lengths", sorted(set(len(s) for s in seqs)))
print("gap or missing characters", gaps, "of", total, "fraction", round(gaps/total, 4))

sequences 17 declared sites 1998
lengths [1998]
gap or missing characters 36 of 33966 fraction 0.0011

- 5 numbers in the answer match no logged tool result: 0.0011, 36, 33966, 95. Run the tool that measures each number, or remove the number, or say how you derived it from logged results.
- The branch support values come from run.contree. The read at #19 returned no file content, so the support values are not visible in the log. They must be checked against the file before use.
Fix these problems. Then give the answer again.

(LngfishAu:0.1709765394,(LngfishSA:0.1884042664,LngfishAf:0.1648765322)100:0.1073673156,(Frog:0.2562938256,((((Turtle:0.2215450915,(Crocodile:0.3057507446,Bird:0.2310620493)97:0.0651606330)70:0.0364894400,Sphenodon:0.3447780527)51:0.0204536475,Lizard:0.3861439885)99:0.0739888076,(((Human:0.1851778377,(Seal:0.0944230016,(Cow:0.0823003161,Whale:0.1012908438)99:0.0404371901)71:0.0252744924)93:0.0340669342,(Mouse:0.0584259465,Rat:0.0905415196)100:0.1218030843)100:0.0607194581,(Platypus:0.1920645910,Opossum:0.1510258428)96:0.0373287871)100:0.1490890925)100:0.1275997766)100:0.0940810002);

path = "{data}/minh2020-iqtree2/example.phy"
with open(path) as f:
    lines = [l.rstrip("\n") for l in f if l.strip()]
header = lines[0].split()
n, L = int(header[0]), int(header[1])
body = lines[1:]
seqs = [""] * n
if len(body) == n:
    for i, l in enumerate(body):
        parts = l.split(None, 1)
        seqs[i] = parts[1].replace(" ", "") if len(parts) > 1 else ""
else:
    for i, l in enumerate(body):
        parts = l.split(None, 1)
        if i < n:
            seqs[i] = parts[1].replace(" ", "") if len(parts) > 1 else ""
        else:
            seqs[i % n] += l.replace(" ", "")
total = sum(len(s) for s in seqs)
gaps = sum(s.count("-") + s.count("?") for s in seqs)
print("sequences:", n, "declared sites:", L)
print("sequence lengths:", sorted(set(len(s) for s in seqs)))
print("total positions:", total)
print("gap or missing characters:", gaps)
print("gap fraction:", round(gaps / total, 4))

sequences: 17 declared sites: 1998
sequence lengths: [1998]
total positions: 33966
gap or missing characters: 36
gap fraction: 0.0011

The final answer

The model wrote this answer. The harness shows it to the scientist after the review.

Settings used, from the decision record: Substitution model: MFP · Bootstrap type: ufboot · Number of bootstrap replicates: 1000 · Random seed: 1.

Checks

Review findings

The review recorded 9 findings. A rule finding comes from a fixed check in the harness. A referee finding comes from a second model that reads the record. The harness shows the findings to the scientist with the final answer. The record does not mark a finding as fixed. Thus a finding from an early review round can apply to a draft that the model corrected later.

Numbers in the answer

The last claim check read 28 numbers in the answer. 23 numbers match a logged result. 5 numbers have no source in the record.

Deviations

The model did not try to change a choice of the scientist.

Failed tool calls

1 tool call failed. The model then tried again or used another tool. The session above shows each failure.

Data integrity

Each data file has the same SHA-256 hash now as at the time of the step that read it. Where the download script (fetch.sh) gives a hash, the file also has that hash. The run did not change the data.

A SHA-256 hash is a fingerprint of the file contents. If one byte of the file changes, the hash changes. The table shows the first 12 characters.

How to repeat it

Get the data. The script downloads the files and checks their SHA-256 hashes where it lists them.

CUVETTE_DATA={data} bash bench/papers/minh2020-iqtree2/fetch.sh

Run the same case with Cuvette. The script gives the same answers from bench/papers/minh2020-iqtree2/bench.yaml.

cuvette bench papers --papers minh2020-iqtree2 --models claude:claude-haiku-5-5

Repeat each step by hand in the program. For each step, the harness records a manual route: the menu path or the code that gives the same result. This list does not include comparison runs.

1. infer_tree (step n1)

   Run: iqtree3 -s <alignment> -m <model> -B <replicates> -seed <seed> -T 1

   • -s

     {data}/minh2020-iqtree2/example.phy

   • -m = MFP
   • -B = 1000
   • -seed = 1
   • Warning: If you keep the default random, you get a different result.

   The manual route that the harness recorded

   /bin/sh {other volume}/tools/overnight/claude-final/catalog/phylo/scripts/iqtree.sh /opt/homebrew/bin/iqtree3 infer --alignment {data}/minh2020-iqtree2/example.phy --model MFP --bootstrap-type ufboot --replicates 1000 --seed 1

   The manual route gives the same numbers. An automatic test in Cuvette checks this.

Figure

Run facts

The session

This is the session as the scientist sees it, in order. Decision cards show the answer that the script gave and where that answer comes from. Each step shows the program, its version, the input file hash and the outputs. Click a line to see more.

- Setup values:
Alignment:
- Alignment method (MAFFT) (algorithm): auto
- Trim the alignment (trim): false
Tree:
- Substitution model (model): MFP
- Bootstrap type (bootstrap_type): ufboot
- Number of bootstrap replicates (replicates): 1000
- Outgroup (outgroup): 
- Random seed (seed): 1
Ask the scientist: Largest gap fraction of a kept column (max_gap_fraction).
Fixed by the lab: Random seed (seed) = 1.
The harness fills these values into your tool calls automatically. If you give a different value, the harness logs a deviation and shows it to the scientist.
- Mode is now watch: you run the steps.

report={work}/infer_tree-1/run.iqtree
seed=1
bootstrap_type=ufboot
bootstrap_replicates=1000
treefile={work}/infer_tree-1/run.treefile
consensus_tree={work}/infer_tree-1/run.contree
n_sequences=17
n_sites=1998
model=TIM2+F+I+G4
best_model_bic=TIM2+F+I+G4
log_likelihood=-21152.5377
tree_length=4.2173
constant_sites=686
site_patterns=1152

- 3 numbers in the answer match no logged tool result: 20261009, 53857, 157. Run the tool that measures each number, or remove the number, or say how you derived it from logged results.
- The log-likelihood must be reported with more decimal places.
Fix these problems. Then give the answer again.

The final answer

The model wrote this answer. The harness shows it to the scientist after the review.

Settings used, from the decision record: Substitution model: MFP · Bootstrap type: ufboot · Number of bootstrap replicates: 1000 · Random seed: 1.

Checks

Review findings

The review recorded 4 findings. A rule finding comes from a fixed check in the harness. A referee finding comes from a second model that reads the record. The harness shows the findings to the scientist with the final answer. The record does not mark a finding as fixed. Thus a finding from an early review round can apply to a draft that the model corrected later.

Numbers in the answer

The last claim check read 5 numbers in the answer. 2 numbers match a logged result. 3 numbers have no source in the record.

Deviations

The model did not try to change a choice of the scientist.

Failed tool calls

1 tool call failed. The model then tried again or used another tool. The session above shows each failure.

Data integrity

Each data file has the same SHA-256 hash now as at the time of the step that read it. Where the download script (fetch.sh) gives a hash, the file also has that hash. The run did not change the data.

A SHA-256 hash is a fingerprint of the file contents. If one byte of the file changes, the hash changes. The table shows the first 12 characters.

How to repeat it

Get the data. The script downloads the files and checks their SHA-256 hashes where it lists them.

CUVETTE_DATA={data} bash bench/papers/minh2020-iqtree2/fetch.sh

Run the same case with Cuvette. The script gives the same answers from bench/papers/minh2020-iqtree2/bench.yaml.

cuvette bench papers --papers minh2020-iqtree2 --models ollama:qwen3:8b

Repeat each step by hand in the program. For each step, the harness records a manual route: the menu path or the code that gives the same result. This list does not include comparison runs.

1. infer_tree (step n1)

   Run: iqtree3 -s <alignment> -m <model> -B <replicates> -seed <seed> -T 1

   • -s

     {data}/minh2020-iqtree2/example.phy

   • -m = MFP
   • -B = 1000
   • -seed = 1
   • Warning: If you keep the default random, you get a different result.

   The manual route that the harness recorded

   /bin/sh {other volume}/tools/overnight/checkout/catalog/phylo/scripts/iqtree.sh /opt/homebrew/bin/iqtree3 infer --alignment {data}/minh2020-iqtree2/example.phy --model MFP --bootstrap-type ufboot --replicates 1000 --seed 1

   The manual route gives the same numbers. An automatic test in Cuvette checks this.

Figure

Run facts