# Esl Compalign

> Use when comparing a test multiple sequence alignment against a trusted reference alignment to compute accuracy. Requires Stockholm format files with

- Skill: `vimalinx/esl-compalign` (Agent Skill, multi-file: 2 files)
- Install (CLI): `npx skillmds add vimalinx/esl-compalign`
- Raw SKILL.md: https://api.skillmd.com/api/skills/vimalinx/esl-compalign/raw
- Safety review: pending
- Works with: Claude Code, Claude.ai, OpenAI Codex
- Category: Coding & Dev Tools
- Author: vimalinx (https://skillmd.com/u/vimalinx)
- Updated: 2026-09-09
- Page: https://skillmd.com/skills/vimalinx/esl-compalign

---


# esl-compalign

## Quick Start
- **Command**: `esl-compalign [-options] <trusted file> <test file>`
- **Local executable**: `/home/vimalinx/miniforge3/envs/bio/bin/esl-compalign`
- **Full reference**: See `references/help.md`

## When To Use This Tool

- Use `esl-compalign` when evaluating a test alignment against a trusted reference alignment.
- It is appropriate for benchmark or method-comparison workflows where both alignments are in Stockholm format with RF annotation.
- Use `-c` when you want per-column statistics and `-p` when comparing alignment accuracy against posterior probability information.
- Reach for it when you need an explicit quantitative comparison of alignment quality rather than just visual inspection.

## Common Patterns

```bash
# Compare a test alignment against a trusted alignment
esl-compalign trusted.sto test.sto

# Emit per-column statistics instead of per-sequence statistics
esl-compalign -c trusted.sto test.sto

# Compare accuracy as a function of posterior probability
esl-compalign -p trusted.sto test.sto

# Write column-wise stats in ssdraw-compatible format
esl-compalign -c --c2dfile compalign.dfile trusted.sto test.sto
```

## Recommended Workflow
1. Prepare trusted and test alignments in Stockholm format with #=GC RF markup
2. Verify sequences appear in identical order in both files
3. Confirm #=GC RF markup has identical number of non-gap characters in both files
4. Run `esl-compalign <trusted file> <test file>` to compute accuracy

## Guardrails
- Both files must be in Stockholm format with #=GC RF markup
- Sequences must occur in the same order in the two files
- Number of non-gap characters in #=GC RF markup must be identical between files
- `-p` only makes sense when posterior probability annotation exists in the tested alignment context