Tools
Liatir offers four kinds of analysis step: built-in tools that work out of the box, native tools that use programs installed on your machine, AI Tools that use local AI Models, and External Workflows that run saved engine projects such as Nextflow.
They all share the same layout, run history, and results view, and they can be mixed freely in a pipeline whenever their inputs and outputs match.
Built-in tools
Bundled with Liatir, so there is nothing to install. They work identically on every machine.
| Tool | Description |
|---|---|
| FastQC | Per-base quality, GC content, adapter detection, duplication levels |
Native tools
Require the corresponding program to be installed on your machine. Liatir checks availability when you open the tool page and shows install instructions (Homebrew, apt, conda) if it is missing.
| Tool | Subcommand | Input formats |
|---|---|---|
| seqkit stats | stats | FASTA, FASTQ (compressed or not) |
| Samtools | flagstat | BAM, SAM, CRAM |
| Samtools faidx | faidx | FASTA, FASTA.GZ |
| BWA-MEM | mem | FASTA + FASTQ |
| Minimap2 | — | FASTA/MMI + FASTQ/FASTA |
| BCFtools | stats | VCF, VCF.GZ, BCF, BCF.GZ |
| BCFtools filter | filter | VCF, VCF.GZ, BCF, BCF.GZ |
| SnpEff | — | VCF, VCF.GZ |
| fastp | — | FASTQ (single or paired-end) |
AI Tools
AI Tools are documented separately because interpreting their results requires model-specific context.
| AI Tool | Description |
|---|---|
| Single-cell Embedding | Foundation-model cell embeddings from .h5ad inputs |
Start with Local AI for bioinformatics if you are new to these outputs.
External Workflows
External Workflows can run on their own or as one reusable node in a Liatir pipeline. The first adapter uses a system-installed Nextflow and Java runtime.
| Engine | Description |
|---|---|
| Nextflow | Saved local or revision-pinned workflows with declared inputs, exact outputs and engine provenance |
Scientific viewers
Viewers inspect output artifacts produced by tools and pipelines.
| Viewer | Description |
|---|---|
| 3D Structure Viewer | PDB/mmCIF/CIF structure inspection |
| Genome Track Viewer | BED and genome-track inspection |
| Single-cell Viewer | profiled AnnData and bounded embedding previews |
Common UI pattern
Every tool page follows the same layout:
- Dependency check — if a required native binary or workflow engine is missing, Liatir explains what is needed.
- Input form — file pickers pre-filtered by compatible extension, plus any tool-specific options.
- Run button — starts the analysis and streams progress or logs when available.
- Results panel — shows parsed stats, tables, text output, charts, and generated files.
- Run history sidebar — all past runs for this tool, selectable to re-display their results.
Result views
Tool results are displayed in a consistent format:
- stats appear as readable key-value grids;
- long text output can be expanded when needed;
- charts are interactive where available;
- generated files appear above the report with actions.
Output files
Some tools produce output files as part of their results. These appear in the results panel above the stats sections, with two actions:
- Add to Data — registers the file in the Data library immediately, making it available as an input to the next step.
- Save as… — opens the system save dialog so you can copy the file to a location of your choice.
Run history
Each tool keeps a persistent run history. Selecting a past run in the sidebar re-renders its output without re-running the tool. Run records include:
- Tool name and version
- Input file paths
- Parsed results
- Output file references
- Timestamp
Pipeline integration
Tools and saved External Workflows expose compatible inputs and outputs to the pipeline builder. This allows an output file from one step to be connected to a compatible input in the next step. See Pipeline Overview for details.