Using WGCNA (weighted gene co-expression network analysis) to identify the hub genes of skin hair follicle development in fetus stage of Inner Mongolia cashmere goat

Abstract

Objective: Mature hair follicles represent an important stage of hair follicle development, which determines the stability of hair follicle structure and its ability to enter the hair cycle. Here, we used weighted gene co-expression network analysis (WGCNA) to identify hub genes of mature skin and hair follicles in Inner Mongolian cashmere goats.

Methods: We used transcriptome sequencing data for the skin of Inner Mongolian cashmere goats from fetal days 45-135 days, and divided the co expressed genes into different modules by WGCNA. Characteristic values were used to screen out modules that were highly expressed in mature skin follicles. Module hub genes were then selected based on the correlation coefficients between the gene and module eigenvalue, gene connectivity, and Gene Ontology (GO)/Kyoto Encyclopedia of Genes and Genomes (KEGG) enrichment analysis. The results were confirmed by quantitative polymerase chain reaction (qPCR).

Results: Ten modules were successfully defined, of which one, with a total of 3166 genes, was selected as a specific module through sample and gene expression pattern analyses. A total of 584 candidate hub genes in the module were screened by the correlation coefficients between the genes and module eigenvalue and gene connectivity. Finally, GO/KEGG functional enrichment analyses detected WNT10A as a key gene in the development and maturation of skin hair follicles in fetal Inner Mongolian cashmere goats. qPCR showed that the expression trends of 13 genes from seven fetal skin samples were consistent with the sequencing results, indicating that the sequencing results were reliable.n.

Fig 1. Schematic diagram of skin hair follicle development.

(A) Primary hair follicles; (B) secondary hair follicles. PC: Placode; DC: Dermal condensate; DP: Dermal papilla; IRS: Inner root sheath; HFSC: Hair follicle stem cells; SG: Sebaceous gland.

Fig 2. Power value curve.

Fig 3. Module hierarchical clustering tree.

Dynamic tree cut indicates module divided according to clustering results; merged dynamic indicates module divided according to similarity of the module. Analysis was carried out according to the merged module. Vertical distance in tree diagrams represents distance between two nodes (between genes); horizontal distance is meaningless.

Fig 4. Heat map of sample expression pattern.

Abscissa represents the sample and ordinate represents the module, based on module eigenvalues. Red: High expression; green: Low expression.

Fig 5. Venn diagram of key candidate genes for specific target modules.

Fig 6. GO-biological process.

**P<0.01. Red box: GO terms directly involved in the development of skin hair follicles.

Fig 7. GO-cellular component.

**P<0.01.

Fig 8. GO-molecular function.

**P<0.01.

Fig 9. KEGG signal pathway enrichment diagram.

Abscissa indicates percentage of genes enriched by each signaling pathway relative to total genes of each signaling pathway; numbers to right of columns indicate number of genes enriched by each signaling pathway; red boxes: Signaling pathways involved in development of skin hair follicles. *P<0.05, **P<0.01.

Fig 10. Gene enrichment map of skin follicle development and GO-biological process.

Large nodes: GO terms directly involved in the biological process of skin hair follicle development; small nodes: Enriched genes; red boxes: Enriched to three GO terms.

Fig 11. Gene enrichment diagram of KEGG signaling pathway in skin hair follicle development.

Nodes with yellow background: Signal pathways; nodes with blue background: Genes; red boxes: Genes enriched in at least three signal pathways.

Fig 12. Venn diagram of GO/KEGG enrichment analysis.

A: The genes marked by the red box are all enriched to 3 GO terms. B: The gene enriched to ≥ 3 signal pathways is marked by the red frame.

Fig 13. Fluorescent QPCR to verify the sequencing results.

Spearman’s rank correlation: 0.8 < |Rs| < 1 extremely strong correlation, 0.6 < |Rs| < 0.8 strong correlation, 0.4 < |Rs| < 0.6 medium correlation, 0.2 < |Rs| < 0.4 weak correlation, 0 < |Rs| < 0.2 no correlation.