1.1 How to start a analysis and which analysis type should I use

On Drylab home page, you will find several option to start your research: Default Analysis, Advanced Analysis, Research. Which one you choose depends on what stage your research is in, how granularly you want to control and interact with the results, and what output you're looking for.

1. Default Analysis

Best for: Everyday coding tasks, quick data exploration, simple plots, and one-off questions.

What it does

• Opens a standard notebook with a conversational AI assistant

• You write prompts, the AI writes and runs code cell by cell

• No automatic planning or multi-step orchestration

• Straightforward request → response workflow

Use when you want to

• Load and inspect a dataset quickly

• Make a single plot or run a quick statistical test

• Ask a simple bioinformatics question

• Prototype code before committing to a full pipeline

Example prompts
Load my CSV and show summary statistics Plot a histogram of the expression column What is the difference between PCA and UMAP? Install scanpy and load my h5ad file

Skill level needed
Beginner-friendly. No prior bioinformatics experience required.

2. Advanced Analysis

Best for: Multi-step scientific analyses, full pipelines, large datasets, and complex workflows requiring tools and databases.

What it does

• Full agentic AI — plans, executes, reviews, and iterates automatically

• Proposes a structured step-by-step plan before executing

• Calls databases, accelerated tools, and pipelines automatically

• Handles errors, retries, and adapts based on intermediate results

• Produces publication-quality figures and reproducible notebooks

Use when you want to

• Run a complete single-cell RNA-seq analysis end to end

• Perform differential expression with GO enrichment and visualization

• Dock multiple drug compounds against a protein target

• Build a machine learning model on biological data

• Run an nf-core pipeline on raw sequencing data

Example prompts

Run a full scRNA-seq analysis on my h5ad file including QC, clustering, annotation, and differential expression between conditions

Predict the structure of 5 protein sequences using Chai-1 and compare their binding pockets

Run nf-core-rnaseq on my samplesheet and perform downstream differential expression analysis

Skill level needed

Intermediate to advanced. You should understand your biological question clearly — the AI handles the technical execution.

3. Research

Best for: Literature review, hypothesis generation, summarizing papers, and exploring the scientific landscape before doing any computation.

What it does

• Searches PubMed, preprint servers, and biological databases

• Summarizes papers, extracts key findings, and identifies research gaps

• Connects findings across multiple sources

• Helps formulate hypotheses and experimental designs

• Can parse uploaded PDFs and DOIs directly

Use when you want to

• Understand the current state of a research field

• Summarize a paper or compare multiple papers

• Find relevant datasets or tools for a biological question

• Generate hypotheses before running experiments

• Get a literature-backed explanation of a gene, pathway, or disease

Example prompts

" Summarize the latest research on KRAS G12D inhibitors in pancreatic cancer What are the key papers on spatial transcriptomics methods published in 2023-2024? Find datasets on GEO related to Alzheimer's single-cell RNA-seq What is the biological role of FOXP3 in regulatory T cells? "

Skill level needed

Beginner-friendly. No coding required.

Which One Should You Start With?