fig3: reduce width

content/05.from-gene-networks-to-machine-learning-derived-gene-modules.md

@@ -12,7 +12,7 @@ The first row illustrates the Pathway-Level Information Extractor (PLIER) method
 This approach generates interpretable gene modules but only captures linear relationships.
 The second row shows a  variational autoencoders (VAEs), which can capture both linear and nonlinear patterns but does not integrate prior knowledge, resulting in less interpretable gene modules.
 The third row highlights interpretable VAEs, which incorporate prior knowledge into its decoder, achieving a balance between capturing complex relationships and maintaining biological interpretability.
-](images/fig3.svg "Omnigenic model"){#fig:fig3 width="100%"}
+](images/fig3.svg "Omnigenic model"){#fig:fig3 width="90%"}
 
 ### From single gene analysis to gene module discovery with machine learning approaches