Tool:KeggExp: a web server for visualizing KEGG pathways maps when analyzing expression profile data

Question

[liux.bio]

Overview

KeggExp is a biologist-friendly web server to integrate KEGG pathways and expression profile data (data generated by Microarray, RNA-seq, or mass spectrometry). It aims to adopt interactive visualization techniques to help researchers find key kegg pathways or genes that are worth further attention in analyzing expression profile data. It has following features:

• We can zoom and drag the pathway map
• We can highlight the genes of interest in the KEGG pathway map
• It displays the pathway genes' expression in a dataset (if a dataset is available) and we can select the co-expressed genes to highlight them in the pathway map
• We can highlight differentially-expressed genes in a differential analysis
• We can dynamically highlight differentially-expressed genes across groups in a dataset

Main features

Zoom and drag the pathway map

---

---

Highlight co-expressed genes in the pathway map

Users could select co-expressed genes in the dendgrogram & heatmap, and highlight them in the pathway map.

---

---

Highlight differentially-expressed genes/proteins in the pathway map

When users click one experimental group, the up-expressed genes will become red and the down-expressed genes will become green.

---

---

Operations for visualizing multiple pathways

One pathway often have cross talks with other pathways. Users can click pathway box in the pathway map and then explore the corresponding pathway. The pathways that have been explored are listed in the left top, users can select one pathway they want to explore.

---

---

Availability

KeggExp is available at http://www.fgvis.com/expressvis/KeggExp. If any questions or suggestions, feel free to contact me.

[Zhilong Jia]

Powerful. I will recommend to my colleagues. Two small bugs. 1. the try samples cannot be downloaded correctly. 2.heatmap cannot show negative value. BTW, how does it calculate the group expression per gene?

For 1. For 2.

[liux.bio]

For the first bug, the browser should be allowed to download multiple files simultaneously.

---

For the second bug, heatmap is designed to not to show negative value. The expression profile data should be processed data, in which the fold change of one gene between group A and group B could be calculated with ValueA/ValueB; that is to say, the expression value should not be transformed to log2 value. PS: you could double click one gene in the heatmap and show this gene's expression with scattermap.

---

As for the group expression per gene, we use the average of values of all the replicates in the group.

---

Many thanks for your feedback. I will add more information to the tutorial.

[Zhilong Jia]

1. try to download it at the deployed website. you will see the downloaded files as follow.

2. The expression profile data should be processed data, in which the fold change of one gene between group A and group B could be calculated with ValueA/ValueB. Actually, it is one group, not group A/group B, per column. In most cases, logging (or voomming in limma) the data will make it follow a normal distribution. A negative value can be used in the form of ValueA/ValueB.

Thank you. KEGGexp is useful.