How to build a maximum likelihood phylogenetic tree with IQ-TREE2 and assess bootstrap support

Question

I have a protein multiple sequence alignment of 80 sequences across multiple species. I want to build a maximum likelihood phylogenetic tree with bootstrap support values. What is the correct IQ-TREE2 command, and how do I choose the substitution model? Should I use standard bootstrap or ultrafast bootstrap?

Admin · Accepted Answer

IQ-TREE2 handles model selection and tree inference together. Here is the standard workflow:

**Basic run with automatic model selection + ultrafast bootstrap (UFBoot)**
```bash
iqtree2 
  -s alignment.fasta 
  -m TEST 
  -bb 1000 
  -alrt 1000 
  -T AUTO 
  --prefix my_tree
```

**Parameters explained:**
- `-m TEST`: automatically tests all substitution models and picks the best by BIC
- `-bb 1000`: ultrafast bootstrap with 1000 replicates (much faster than standard)
- `-alrt 1000`: SH-aLRT test as an alternative support metric
- `-T AUTO`: automatically detect optimal CPU threads

**For protein sequences, use a protein model set:**
```bash
iqtree2 
  -s protein_alignment.fasta 
  -m TEST 
  -mset LG,WAG,JTT,Dayhoff 
  -bb 1000 -alrt 1000 
  -T 8 --prefix protein_tree
```

**Standard bootstrap (slower, but more conservative):**
```bash
iqtree2 -s alignment.fasta -m MFP -b 100 -T 8
```

**Output files:**
- `.treefile`: ML tree with support values (use this for visualization)
- `.iqtree`: full analysis log with model details
- `.contree`: consensus tree

**Interpreting support values:**
- UFBoot ≥ 95 ≈ reliable (equivalent to standard bootstrap ≥ 70)
- aLRT ≥ 80 is considered significant
- Report both: `node_label/aLRT/UFBoot` convention

Visualize with FigTree (free) or iTOL (web-based).