How to customize analysis

Once you've selected your datasets in the previous step, get ready to customize the analysis steps. While we do choose default parameters for each step, you have the ability to change them at any time and reanalyse datasets.

The specific analysis steps depend on the type of dataset and the parameters chosen (checkout the disabled sections in the screenshot below for a scRNA-seq dataset). Whether it's batch correction or multi-modal analysis, we've got you covered!

The analysis pipeline follows the standard analysis workflows described in Orchestrating Single-Cell Analysis:

• Removal of Low-Quality Cells: Bid farewell to low-quality cells as we whisk them away from your dataset.
• Normalization and Log-transformation: Transform your data into a magical realm of normalized and log-transformed wonder.
• Modeling the Mean-Variance Trend Across Genes: Unveil the secrets of gene behavior with our meticulous modeling of the mean-variance trend.
• Principal Components Analysis on Highly Variable Genes: Delve into the essence of your data using a principal components analysis, revealing its core elements.
• Clustering with Graph-Based Methods: Watch as we weave a web of clusters, connecting similar cells in your dataset through graph-based methods.
• Usual Dimensionality Reductions (t-SNE/UMAP): Witness the unveiling of the underlying patterns as we employ t-SNE and UMAP for dimensionality reduction.
• Marker Detection for each cluster: Uncover the distinct marks of each cluster as we perform marker detection, revealing their unique traits.
• Gene Set Enrichment for each cluster: where we unveil the hidden treasures among the top markers for each cluster!
• Cell Type Annotation for each cluster: Through the power of our magic wand, we annotate each cluster with cell types, drawing from user-selected reference datasets.

While we won't delve into the minute details here, don't worry! The interface provides all the information you need about available parameters and their descriptions.