Galaxy RNA-seq Analysis Tutorial

Introduction

Analyzing biological data can seem difficult when you do not know how to program. For example, a biology student may have sequencing data from healthy and diseased plants but may not know how to process millions of DNA or RNA reads. This is where Galaxy bioinformatics can help.

Galaxy is a free, web-based platform that allows researchers, students and teachers to analyse biological data through a graphical interface. It provides access to bioinformatics tools for sequence quality control, read mapping, gene expression analysis and data visualisation without requiring users to write code for every step.

In this Galaxy RNA-seq analysis tutorial, you will learn how to upload FASTQ files, check sequencing quality with FastQC and MultiQC, map reads using HISAT2 or STAR, count gene reads and perform differential gene expression analysis using DESeq2. You will also learn how to create reusable workflows and share your results.

Whether you are a biology student, a beginner in bioinformatics or a researcher learning RNA-seq analysis, this step-by-step guide will help you understand the main stages of the process.

Note: Galaxy reduces the need to write code, but you still need to understand experimental design, data quality and statistical results. Available tools, computing limits and file-handling options can vary between Galaxy servers.

1. The Galaxy Interface

When you open Galaxy (for example, usegalaxy.org) and log in, the screen is divided into simple parts:

• Tool panel (left): A list of tools, grouped by topic such as Get Data, FASTQ Quality Control, Mapping and RNA-seq. A search box at the top helps you find any tool quickly.
• Centre panel: This is your workspace. When you click a tool, its settings open here. You choose your input, adjust options and press Run.
• History panel (right): This is your lab notebook. Every file you upload and every result you create appears here as a small box. Green generally indicates successful completion, yellow indicates a queued or running job, and red indicates a failed job.

Example: Think of Galaxy like a kitchen. The tool panel is your cupboard of equipment, the centre panel is your cooking counter, and the history panel is the plate where finished dishes are placed.

A good habit is to give your history a clear name, such as "Rice drought RNA-seq", so you can find it months later.

2. Uploading Data

Every analysis starts with data. In Galaxy you click Upload Data at the top of the tool panel. You have several choices:

1. Upload from your computer – drag and drop your files.
2. Paste a link (URL) – useful for big public datasets, because Galaxy downloads them directly.
3. Shared data libraries – many teachers and labs keep practice files ready for you.

Galaxy needs to know what kind of file it is. Common file formats include FASTQ for sequencing reads, FASTA for DNA, RNA or protein sequences, and GTF for gene annotations. Galaxy may infer the file format automatically, but you should always verify it.

Example: Priya has two sets of reads, control_1.fastq and drought_1.fastq. She uploads both, renames them clearly, and sees two green boxes in her history. Her data is ready.

Tip: If your data are paired-end (two files per sample), organise them as a paired collection. This keeps files together and saves a lot of confusion later.

3. Quality Control (QC)

Raw sequencing data is never perfect. It can contain low-quality bases, leftover adapter sequences, or very short reads. If we ignore this, our final results may be wrong. This is like washing vegetables before cooking.

Step 1 – Check with FastQC. This tool reads your file and gives a report with graphs. The most useful ones are:

• Per-base sequence quality: This graph shows the quality of base calls at each position in the reads. Higher Phred scores indicate more reliable base calls. Q30 corresponds to an estimated base-calling error probability of 0.1%. Quality should be interpreted according to the sequencing platform and experiment.
• Adapter content: Is there leftover technical sequence?
• Sequence length: Are reads the expected size?

Step 2 – Combine with MultiQC. If you have many samples, MultiQC puts all FastQC reports into one clean page, so you can compare them at a glance.

Step 3 – Clean with Trimmomatic or fastp. These tools can remove adapter sequences and low-quality bases when trimming is justified by the quality-control results.

Step 4 – Check again. Run FastQC again on the cleaned data to assess whether trimming improved the identified quality issues.

Example: Priya's FastQC report shows quality dropping at the end of her reads. After trimming, the graph stays in the green zone. She now trusts her data.

4. Mapping

Now we ask: where does each read come from in the genome? This step is called mapping or alignment. Think of it as a jigsaw puzzle. The genome is the picture on the box, and each read is a tiny piece that must be placed in the correct position.

For RNA-seq, we use a splice-aware mapper, because RNA comes from genes that are cut and joined (spliced). Popular choices in Galaxy are:

• HISAT2 – A splice-aware aligner designed for efficient alignment of RNA-seq reads.
• STAR – A splice-aware aligner known for its high alignment speed, particularly for large RNA-seq datasets.

You provide three things: the cleaned reads, a reference genome (built-in options exist for many species), and options such as strandedness. The alignment output is commonly stored in a SAM or BAM file, which records mapped positions and alignment information for sequencing reads. Unmapped reads may also be represented in the alignment file.

After mapping, examine the alignment summary and compare the results with expectations for your organism, library type, and experiment. A low mapping rate may indicate poor read quality, contamination, an incorrect reference genome, or differences between the sample and reference. There is no single mapping percentage suitable for every experiment.

5. RNA-seq Workflow Without Code

RNA-seq tells us which genes are switched on or off in different conditions, for example healthy versus diseased, or normal versus drought. The full workflow in Galaxy is a chain of simple steps:

1. Upload the sequencing reads and select a reference genome and compatible gene annotation.
2. QC with FastQC and MultiQC.
3. Trim with fastp or Trimmomatic.
4. Map with HISAT2 or STAR.
5. Count reads assigned to genes using featureCounts or HTSeq-count, together with a compatible gene annotation. The resulting count matrix typically contains genes as rows and samples as columns.
6. Perform differential expression analysis using DESeq2 or edgeR with an appropriate gene-count matrix. Define the experimental groups and statistical design, accounting for relevant factors such as treatment and batch when necessary. Biological replicates are important for estimating variability and drawing reliable conclusions.
7. Visualise with volcano plots, heatmaps and PCA plots.

Each step feeds the next, and the highlighted step gives the answer most people are looking for.

The key output is a table listing each gene, its log2 fold change, and its adjusted p-value. The log2 fold change indicates the direction and magnitude of the expression difference, while the adjusted p-value accounts for multiple statistical testing. Genes with an adjusted p-value below a predefined threshold, commonly 0.05, are considered statistically significant. Their biological importance should also be evaluated using effect sizes and other biological evidence.

Example: Priya finds that a gene linked to water-stress tolerance is strongly increased in drought plants. That one result may guide her next lab experiment, and she reached it with only mouse clicks.

6. Sharing Workflows

Science becomes powerful when it can be repeated and shared. Galaxy makes this easy in two ways.

Create a workflow. After finishing an analysis, choose Extract Workflow from your history. Galaxy converts your steps into a reusable pipeline. Next time, you only upload new data and press Run. The same configured steps can be applied to new datasets, helping make the analysis consistent and reproducible.

Share it. You can:

• Share a history or workflow with a link
• Share with specific people by their email
• Publish it so everyone can find it
• Export it to a public workflow hub such as WorkflowHub

This supports reproducibility: another scientist can open your history, see every tool and setting, and repeat your work. It is also wonderful for teaching, because a teacher can share one workflow with a whole class.

Example: Priya shares her workflow with her classmate Arjun. He uploads his own rice samples and runs the same workflow, allowing him to repeat the analysis more efficiently.

Conclusion

Galaxy proves that powerful bioinformatics is not only for programmers. With a browser, you can upload data, clean it, map it, count genes, find differences and share your whole method. Every step is visible, recorded and repeatable.

For a beginner, the best way to learn is to practise: pick a small public dataset, follow the steps once, and then try changing one setting to see what happens. Start simple, stay curious, and let the data tell its story. Coding can come later; good science can start today.