Shulou(Shulou.com)06/01 Report--

This article mainly introduces "how to mine potential co-expression genes in WGCNA". In daily operation, I believe many people have doubts about how to mine potential co-expression genes in WGCNA. The editor consulted all kinds of data and sorted out simple and easy-to-use operation methods. I hope it will be helpful to answer the doubts about "how to mine potential co-expression genes in WGCNA". Next, please follow the editor to study!

Co-expression genes refer to the collection of genes whose expression tends to change cooperatively. It is generally believed that these genes participate in the same biological process, such as participating in the same metabolic pathway, precisely because of functional synergism, resulting in a high correlation in expression.

In WGCNA, the traditional correlation coefficient is multiplied, and the final value is used to characterize the correlation between genes. After calculating such correlation statistics, how to determine which genes are co-expressed?

The method of WGCNA is cluster analysis, which belongs to an unsupervised machine learning algorithm. Through the cluster tree, we can observe which genes belong to the same branch in the cluster tree, and the genes belonging to the same branch can be classified into the same category. In practice, taking into account the large number of genes and so on, we certainly need the algorithm to automatically classify. WGCNA uses the R packet dynamicTreeCut.

For the clustering algorithm, the distance matrix between genes needs to be inputted, first of all, the adjacency matrix between genes needs to be transformed into a distance matrix, and the adjacency matrix can be calculated by multiplying the correlation coefficients. But this value essentially reflects the similarity between genes, not distance. When calculating the distance matrix, WGCNA uses the TOM statistic, which can represent the similarity of nodes in the network. The calculation formula is as follows.

For the two genes I and j, a represents the corresponding value in the adjacency matrix of the two genes, which is the multiplier of the correlation coefficient, and K represents the connectivity of each gene. The formula is as follows.

For the weighted network, it is the sum of the corresponding values of the edges of the node. for example, in the network, gene An is connected to three genes, and the connection degree of gene An is the sum of the values of the corresponding three sides. The l value between two genes represents the sum of the weight product of all sides of the two genes. The formula is as follows

The formula only helps us to understand the process of calculation. In fact, we only need to understand that TOM represents the similarity of nodes. What we want is the distance, so we can use 1 to subtract the similarity directly. The formula is as follows

With the help of TOM value, the correlation coefficient between genes is converted into distance, and then the distance matrix can be used for clustering. The above calculation methods all have corresponding formulas in WGCNA, and the code is as follows

# determine the best value powers for power in multiplicative operation

Welcome to subscribe "Shulou Technology Information " to get latest news, interesting things and hot topics in the IT industry, and controls the hottest and latest Internet news, technology news and IT industry trends.