# Amino Acid Composition

> Use when counting amino-acid letters in raw protein sequence lines inside simple EDirect text pipelines.

- Skill: `vimalinx/amino-acid-composition` (Agent Skill, multi-file: 2 files)
- Install (CLI): `npx skillmds@latest add vimalinx/amino-acid-composition`
- Raw SKILL.md: https://api.skillmd.com/api/skills/vimalinx/amino-acid-composition/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/amino-acid-composition

---


# amino-acid-composition

CLI tool from the Entrez Direct (EDirect) package for computing amino acid composition of protein sequences.

## Quick Start

- **Command:** `printf 'ACDE\n' | PATH=/home/vimalinx/miniforge3/envs/bio/bin:$PATH /home/vimalinx/miniforge3/envs/bio/bin/amino-acid-composition`
- **Local executable:** `/home/vimalinx/miniforge3/envs/bio/bin/amino-acid-composition`
- **Full reference:** See [references/help.md](references/help.md) for complete options and examples

## When To Use This Tool

- Count amino-acid letters in one raw sequence line at a time.
- Emit a fixed 26-row three-letter abbreviation table for each input line.
- Do quick composition checks in lightweight EDirect or shell pipelines where a full FASTA parser would be overkill.

## Common Patterns

```bash
# 1) Count amino acids in a single raw sequence line
printf 'ACDE\n' | \
  PATH=/home/vimalinx/miniforge3/envs/bio/bin:$PATH \
  /home/vimalinx/miniforge3/envs/bio/bin/amino-acid-composition
```

```bash
# 2) Strip FASTA headers first if your input came from a FASTA file
grep -v '^>' proteins.fa | \
  PATH=/home/vimalinx/miniforge3/envs/bio/bin:$PATH \
  /home/vimalinx/miniforge3/envs/bio/bin/amino-acid-composition
```

## Recommended Workflow

1. Feed only raw sequence lines on stdin.
2. Remove FASTA headers and any non-sequence metadata before running the tool.
3. Interpret each 26-row block as the composition for one input line.
4. Aggregate across lines separately if you need per-file rather than per-line totals.

## Guardrails

- This script does not parse FASTA records. A header line such as `>p1` will be treated as sequence text and counted.
- Non-letter characters are stripped, output is case-insensitive, and each input line is processed independently.
- The wrapper depends on the EDirect helper `sort-uniq-count`, so keep the bio / EDirect bin directory on `PATH`.
- Output covers all 26 alphabet letters via three-letter labels (`Ala`, `Asx`, `Xle`, `Pyl`, `Sec`, `Xxx`, `Glx`, etc.), not just the canonical 20 amino acids.