diff --git a/README.md b/README.md
index 026c7db..0c78a3d 100644
--- a/README.md
+++ b/README.md
@@ -1,7 +1,8 @@
# LEGEND: An integrative algorithm for identifying co-expressed and cofunctional genes in multimodal transcriptomic sequencing data
We present a novel method called mu**L**timodal co-**E**xpressed **GE**nes fin**D**er (LEGEND) that performs integrated gene clustering on scRNA-seq and SRT data to identify co-expressed genes at both the cell type and tissue domain levels. LEGEND performs a hierarchical gene clustering with the aim of maximizing intra-cluster redundancy and inter-cluster complementarity.
-![image](docs/assets/img/workflow.jpg)
+
+![image](docs/assets/img/workflow.png)
## Dependencies
diff --git a/docs/_config.yml b/docs/_config.yml
index 63df9d1..8ddd0db 100644
--- a/docs/_config.yml
+++ b/docs/_config.yml
@@ -1,5 +1,5 @@
title: LEGEND
-logo: img/logo.png
+logo: assets/img/logo.png
description: An integrative algorithm for identifying co-expressed and cofunctional genes in multi-omics.
show_downloads: false
google_analytics:
diff --git a/docs/index.md b/docs/index.md
index 8851f21..5ffb35f 100644
--- a/docs/index.md
+++ b/docs/index.md
@@ -20,7 +20,7 @@ toc: true
Exploring co-expressed genes is essential for revealing biologically co-functional genes. However, existing methods for this purpose rely solely on sc/snRNA-seq or spatially-resolved transcriptomic (SRT) data, leading to weaker co-functionality among identified genes. We present `LEGEND` (mu**L**timodal co-**E**xpressed **GE**nes fin**D**er), a novel method that performs integrated gene clustering on sc/snRNA-seq and SRT data for identifying genes co-expressed at both the cell type and tissue domain levels.
-
+
The above figure illustrates the workflow of `LEGEND`. Under the framework of information theory, `LEGEND` estimates gene relevance, redundancy and complementarity in both SRT and sc/snRNA-seq datasets in a pseudo-semi-supervised manner. This information is used to construct a gene-gene redundancy graph, on which hierarchical gene clustering is performed using relative redundancy index (RRI) between neighboring gene nodes. The resulting clusters contain genes that are co-expressed at both tissue domain and cell type levels, suggesting a higher likelihood of biological co-functionality.
@@ -299,6 +299,6 @@ Both genes belong to cluster 48. We can now proceed to plot their spatial expres
sq.pl.spatial_scatter(adata_st, layer='STAGATE_ReX', color=["Caly", "Zcchc18"], figsize=(5, 5))
```
-
+
Areas with high expression of one gene correlate with high expression of the other, and similarly for low expression areas. This illustrates the effectiveness of `LEGEND` in identifying biologically relevant gene clusters.