# Wgsim Eval Pl

> Use when evaluating wgsim simulation results or analyzing simulated read data as part of samtools bioconda installations.

- Skill: `vimalinx/wgsim-eval-pl` (Agent Skill, multi-file: 2 files)
- Install (CLI): `npx skillmds add vimalinx/wgsim-eval-pl`
- Raw SKILL.md: https://api.skillmd.com/api/skills/vimalinx/wgsim-eval-pl/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/wgsim-eval-pl

---


# wgsim-eval-pl

## Quick Start
- **Command:** `wgsim_eval.pl <command> [arguments]`
- **Local executable:** `/home/vimalinx/miniforge3/envs/bio/bin/wgsim_eval.pl`
- **Reference:** See [references/help.md](references/help.md) for detailed usage

## When To Use This Tool

- Evaluate alignments or variant calls produced from `wgsim`-simulated data.
- Compare aligner accuracy on SAM files whose read names encode wgsim truth.
- Post-process SAM files to keep top-scoring hits before comparison.
- Compare two unique-hit alignment outputs from competing mapping workflows.

## Common Patterns

```bash
# 1) Summarize mapping accuracy from a wgsim-derived SAM file
wgsim_eval.pl \
  alneval \
  simulated.sam
```

```bash
# 2) Emit ROC-style alignment evaluation output
wgsim_eval.pl \
  alneval \
  -a \
  simulated.sam
```

```bash
# 3) Evaluate variant calls against wgsim SNP truth
wgsim_eval.pl \
  vareval \
  truth.wgsim.snp \
  calls.pileup.flt
```

```bash
# 4) Keep only the top-scoring SAM hit per read before downstream comparison
wgsim_eval.pl \
  unique \
  alignments.sam
```

## Recommended Workflow

1. Decide which task you need: `alneval`, `vareval`, `unique`, or `uniqcmp`.
2. Confirm the inputs really come from `wgsim`-style simulations, because several subcommands depend on truth encoded in the read names or SNP truth files.
3. Run the chosen subcommand and capture the tabular output for downstream plotting or benchmarking.
4. If necessary, use `unique` before `alneval` or `uniqcmp` so multi-hit alignments are normalized consistently.

## Guardrails

- The first argument must be a subcommand; `--help` and `--version` are not recognized as top-level actions.
- `alneval` and `unique` expect SAM alignments whose read names were generated by `wgsim`; generic SAM files can produce warnings or meaningless accuracy numbers.
- `vareval` expects a `wgsim.snp` truth file plus a pileup-style variant call file.
- `unique` and `uniqcmp` are alignment post-processing/evaluation helpers, not generic SAM validation commands.