Tools
Liatir offers four kinds of analysis step: built-in tools that work out of the box, native tools — real bioinformatics programs, shipped with Liatir — 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
These are real bioinformatics programs, and Liatir ships them. You do not install samtools, bwa, minimap2, bcftools, seqkit, fastp, simpleaf or alevin-fry yourself: they arrive with the application, at the exact versions it was tested with, and they update when Liatir updates. The tool page shows Included with Liatir and the version it is running.
SnpEff and SnpSift are the exception: they are Java programs. Liatir downloads and verifies both together only when you need them, then manages their versions and SnpEff databases. Java 21 or newer must still be installed on your machine.
On Windows these tools run inside WSL2 — see Windows and WSL2 below.
| 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 |
| SnpSift Filter | filter | VCF, VCF.GZ |
| fastp | — | FASTQ (single or paired-end) |
| Single-cell Reference Index | simpleaf index | Genome FASTA + GTF/GFF3 |
| Single-cell Quantification | simpleaf quant | Single-cell index + paired FASTQ |
The two single-cell steps are one toolchain: simpleaf drives piscem for mapping and alevin-fry for counting, so you run one step instead of five. The result is an .h5ad count matrix, which is the input the Single-cell Embedding AI Tool and the Single-cell Viewer expect.
Common single-cell references can be selected from a ready-made list. Liatir downloads each index once, verifies it and keeps it for reuse; the reference index tool remains available for species or annotation releases not in that list.
Windows and WSL2
Almost none of this software has a Windows build. samtools, bcftools, bwa, minimap2, fastp, simpleaf and alevin-fry are written for Unix and rely on operating-system facilities Windows does not provide; their authors publish Linux and macOS releases only. This is not a Liatir limitation and no Windows version exists to package.
So Liatir on Windows ships the Linux tools and runs them through WSL2, the Windows Subsystem for Linux — the same mechanism it already uses for Nextflow. You still work entirely in the Liatir window; your files stay where they are, and Liatir translates their locations for the tool.
WSL2 must be installed. It is a supported Windows feature, not third-party software: open PowerShell as administrator, run wsl --install, and restart when asked. You do not need to install anything inside it for these tools — Liatir brings its own copy and sets it up for you. Nextflow is different and does need to be installed inside WSL2 yourself.
Liatir unpacks its tools the first time it starts after being installed or updated. It takes a second or two, happens in the background, and does not repeat.
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
What a run writes down
A run records everything it did, not only what it was asked for.
Every file it produced. Each run has its own folder, and everything the tool wrote goes in it — the result, plus whatever it produced along the way: fastp's quality report, the log and execution report a Nextflow workflow leaves behind, working files. The results panel lists the results and tells you how many other files there are; Open run folder shows them in Finder or Explorer. It is there on every screen that displays a run — at the top of the Results screen, beside Explain and Export HTML, and next to View log under the run on the tool page, AI Tools and External Workflows — including for a run that failed, which is when you are most likely to want it.
The same is true of a step inside a pipeline: it is a run like any other and has its own folder. Where you find your results does not change — they still appear under Results in your Data library, grouped by tool.
This matters for a specific reason: some files are written next to your files, under your own filenames — the index bwa builds beside your reference the first time you align against it. A tool that quietly creates five files in your folder and never says so is a tool you cannot audit, so those are recorded with the run too.
The complete log. The run keeps everything the program printed, including its error output, marked as such. It is the full record, not the trimmed version shown while the run is in progress, and it is kept for as long as the run itself — so a result from two months ago can still be explained.
Nothing is deleted for you. Liatir never removes a run on its own, however long the history gets, because a run holds files you made and may still be using. When you want the space back, the Results screen offers an action that deletes older runs — it tells you first how many files go with them and how many of those are in your Data library.
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.