weighted gene co-expression network analysis (weighted gene co-expression netwoek analysis, WGCNA) is a bioinformatics analysis method used to describe gene expression correlation patterns between different samples. It can mine gene modules that are highly correlated at expression levels. And analyze the relationship between the module and specific traits or phenotypes. Search for the keyword "WGCNA" on Pubmed , and you can find that more than 600 related articles have been published in 2021, which shows that its popularity is still not less than that of the year. So today, I will share with you an article that only made WGCNA and sent to 5 points +, titled "Gene correlation network analysis to identify regulatory factors in sepsis" (Journal of Translational Medicine; impact factor 5.531), hurry up Learn to rub a wave of hot spots!
analysis process:
analysis result
1. Sample statistics
p0s GSEpan contains a total of 2686 patients with sepsis in the data set. Among them, 4 samples were excluded due to abnormal data, and the remaining 682 patients were included in the study. According to the infection site, they were divided into 192 cases of pneumonia sepsis, 48 cases of abdominal sepsis, and 442 cases of other sepsis. There were no significant differences in age, sex, and mortality among the three types of sepsis patients (Table 1).
2. Consistent WGCNA and module identification
The study obtained 27 consistent sepsis modules through WGCNA (Fig. 1).Subsequent studies performed gene differential expression analysis on healthy and sepsis samples, and defined up- and down-regulated modules through the significance T value and the median of the difference multiples. For example, the black module was significantly up-regulated in sepsis samples ( Fig. 2a-b). In addition, genes such as ARG1, CD177, and MMP8 were also found to be significantly up-regulated in patients with sepsis, and the difference was the largest (Fig. 2c).
then studied and identified the modules of specific sepsis, namely pneumonia sepsis module, abdominal sepsis module and other sepsis modules, and performed the module preservation analysis, and the results showed specific sepsis The module has better module preservation and result migration (Fig.3). Subsequent research further explored whether the pneumonia-specific sepsis module can also be identified in the consistency module. The results show that each pneumonia sepsis module has at least one consistency module corresponding to it (Fig. 4). For example, the red pneumonia module corresponds to the black consistency module, and there is an overlap of 275 genes, indicating that there may be a common biological pathway between these modules.
Fig.1
Fig.2
Fig.3
strong
Fig.4 _p5img strong empysis
0 Analyze
Then, by calculating the Pearson correlation coefficient between the expression of the characteristic gene of the module and the clinical characteristic, the relationship between the consistency module and the clinical trait was further explored. The results show that the black module is significantly negatively correlated with the mortality rate, that is, the higher the module expression value, the lower the mortality rate; while the light yellow module is significantly positively correlated with the mortality rate (Fig.5).
Fig.5
4. Module function research
Considering the correlation coefficient between black and light yellow modules and mortality is the largest,This study further carried out a biological function enrichment analysis on these two modules. The results showed that the biological functions of myeloid leukocyte-mediated immunity, leukocyte degranulation and neutrophil -mediated immunity share a large number of characteristic genes of the black module (Fig. 6); while the light yellow module is mainly enriched in transformation , Biological processes such as the decomposition process of basic compounds, the decomposition process of heterocycles, and the decomposition process of cellular nitrogen compounds (Fig.7).
Fig.6
Fig.7
5. transcription factor enrichment analysis
is guessed that they are shared by the transcription module of. Factors and other common mechanisms are regulated, so the study conducted enrichment analysis on key modules of transcription factors. It was found that a total of 93 genes in the black module were enriched in the dbcorrdb_CEBPB_ENCSR000BQI_1_m1 motif, with a normalized enrichment score of 5.53, and this motif was directly annotated into the transcription factor CEBPB, so the transcription factor CEBPB was considered black The main transcription regulator of the module. In the light yellow module, a total of 12 genes are enriched in the taipal_etv6_full_ccggaascggaagtn_repr and cisbp_M5425 motifs and are annotated to the transcription factor ETV6. Therefore, ETV6 is the main transcriptional regulator of the light yellow module.
6. Prediction of the interaction between miRNA and the target gene
Then the study determined the verified interaction of miRNA-module characteristic genes from the mirtarbase database,Among them, 1981 miRNAs are potential regulators of the black module. The top 5 miRNAs with the largest number of regulated genes are hsa-miR-335-5p (n=59), hsa-miR-26b-5p (n=57), has -miR-16-5p (n=44), hsa-miR-17-5p (n=42) and has-miR-124-3p (n=38). The 893 miRNAs are considered as potential regulators of the light yellow module, and the top 5 miRNAs that regulate the number of genes are miR-16-5p (n=14), hsa-miR-92a-3p (n=12), hsa -miR-26b-5p (n=9), hsa-miR-615-3p (n=9) and hsa-let-7b-5p (n=8).
7. Survival analysis
In order to further verify the correlation between key modules that are significantly related to mortality and survival results, the study uses the inverse graph embedding (DDRTree) method to integrate high-dimensional space Reduce to a low-dimensional space to preserve the inherent structure of the sample data, and use the K-means method to cluster the samples. The results show that in the two-dimensional space of the black module after dimensionality reduction, the sample is divided into two clusters . The expressions of the characteristic genes of the modules in the two clusters are significantly different, and the KM results confirm the two clusters There are also significant differences in the survival rate of the samples (Fig. 8). Subsequent studies performed gene differential expression analysis on the samples in these two clusters, and these genes were significantly enriched in neutrophil degranulation, neutrophil activation involved in immune response, neutrophil activation and neutralization. In terms of neutrophil-mediated immunity and other biological functions, cluster 2 has more activation functions, including myeloid cell activation, cell activation, leukocyte activation, and bone marrow leukocyte activation (Fig. 9).
After dimensionality reduction and clustering of the light yellow module, the sample is also divided into two clusters in the two-dimensional space.The expression of the characteristic gene of the module can also identify these two clusters, and the survival probability of cluster 2 is significantly lower than that of cluster 1 (Fig. 10). The results further confirmed the correlation between the light yellow module and the patient's death outcome.
Fig.8
Fig.9
Fig.10
+ Highlight analysis:
5 samples were used in this study. The editor summed up the following key points for everyone: 1) Deeply dig into public databases, the data set is not very precise, the data set of this study contains 802 samples, even after screening and elimination, the effective sample size is still very large; 2) Make good use of different types of samples in the data set to increase the diversity of the analysis. For example, this study analyzes both the consistency module and the specific module, and finally the overall and classification are correlated to construct a complete analysis line; 3 ) Grasp one point for longitudinal in-depth analysis. For example, the main focus of this study is the consistency modules that are significantly related to mortality. Then the two modules are analyzed for biological function enrichment, transcription factor enrichment analysis, and miRNA Control prediction, survival analysis, etc., increase the depth of the article's analysis. I hope that through the study of this article, children's shoes can gain something and make a breakthrough.Welcome everyone to private message or leave a message to discuss~
Original link:
Zhang Z, Chen L, Xu P, Xing L, Hong Y, Chen P. Gene correlation network analysis to identify regulatory factors in sepsis . J Transl Med. 2020 Oct 8;18(1):381.
https://www.ncbi.nlm.nih.gov/pmc/articles/PMC7545567/
.
