Manual mode

Manual Submission is the lowest-friction usage path: no command line and no Agent — you pick a tool on the Analysis Tools page, fill in the form, and submit the task.

Typical flow

Tool catalog → select tool → fill parameters → choose files → submit task → inspect logs → download results

1. Browse the tool catalog

Open the Analysis Tools page; tools are grouped by domain:

• Genomics: sequence alignment, variant calling, gene annotation…
• Transcriptomics: differential expression, co-expression, alternative splicing…
• Epigenomics: peak calling, methylation analysis, chromatin accessibility…
• Proteomics: sequence analysis, structure prediction, interaction networks…
• Plotting & visualization: volcano plots, heatmaps, enrichment bubble plots, network graphs…
• Database queries: cross-database ID mapping, gene functional annotation, regulatory evidence retrieval…

Alongside the large set of single-step tools, the catalog also carries two kinds of “multi-step” entries:

Native workflows

The platform ships 5 native multi-step workflows that bundle the stages of a standardized pipeline into a single submission:

Verified direct submissions expand into a task chain: you fill the form once and the platform automatically creates the parent task and per-stage subtasks in order — the Tasks page shows progress for the whole chain. GWAS uses a dedicated one-page frontend form: fill it once, pass the intent review, and the same full task chain is generated.

See Native workflows for details.

One-click pipeline tools

Gene family analysis, homologous gene identification, synteny analysis, and more are also packaged as one-click pipeline tools: fill in species sequences and annotations once, and the platform automatically chains the search, identification, motif, and phylogenetic tree stages. See Homologous genes & gene families.

2. Open the tool detail page and fill the parameter form

Click a tool to open its detail page, where command-line parameters are rendered as a fillable form:

• Required fields are marked with a red asterisk *;
• Hover over a parameter name for its full description;
• Forms auto-validate ranges and enums.

3. Select input files

File parameters are chosen from File Management:

• Upload: drag-and-drop or click to upload; single files up to 100 GB with chunked transfer;
• Formats: FASTA, FASTQ, BAM/SAM, BED, GFF/GTF, VCF, CSV/TSV, TXT, and other common formats;
• Preview: after upload you can preview the first lines of text files online.

Tip: upload files before opening the tool form — the file picker automatically lists everything you’ve uploaded.

4. Submit the task and track it

Fill in the parameters and click Submit. A single-step tool creates one task; a native workflow creates a task chain (a parent task plus per-stage subtasks). On the Tasks page you can:

Task status flow: queued → running → success / failed. For failed tasks, check the error log to locate the cause, adjust the parameters, and resubmit.

5. Provenance and reproducibility

Every task records:

• Parameter snapshot: the complete parameter values at submission time, guaranteeing reproducibility;
• Execution logs: full stdout/stderr;
• Artifact list: all output files and their paths;
• Tool version: the version of each tool used.

Tasks submitted by the Agent leave the same records — any task, from any path, is traceable.

How this relates to the Agent

The tools you call manually are the same tool ecosystem the Agent reaches via MCP. The differences:

Whenever manual is inconvenient, switch seamlessly to Agent-assisted mode.

Next steps

• Native workflows: an end-to-end RNA-seq walkthrough.
• GWAS end-to-end: multi-step genetic analysis submitted from a one-page form.
• Use cases: end-to-end examples to follow.