Silhouette Scores for Arbitrary Defined Groups in Gene Expression Data and Insights into Differential Expression Results

Abstract

BackgroundHierarchical Sample clustering (HSC) is widely performed to examine associations within expression data obtained from microarrays and RNA sequencing (RNA-seq). Researchers have investigated the HSC results with several possible criteria for grouping (e.g., sex, age, and disease types). However, the evaluation of arbitrary defined groups still counts in subjective visual inspection.ResultsTo objectively evaluate the degree of separation between groups of interest in the HSC dendrogram, we propose to use Silhouette scores. Silhouettes was originally developed as a graphical aid for the validation of data clusters. It provides a measure of how well a sample is classified when it was assigned to a cluster by according to both the tightness of the clusters and the separation between them. It ranges from 1.0 to − 1.0, and a larger value for the average silhouette (AS) over all samples to be analyzed indicates a higher degree of cluster separation. The basic idea to use an AS is to replace the term cluster by group when calculating the scores. We investigated the validity of this score using simulated and real data designed for differential expression (DE) analysis. We found that larger (or smaller) AS values agreed well with both higher (or lower) degrees of separation between different groups and higher percentages of differentially expressed genes (PDEG). We also found that the AS values were generally independent on the number of replicates (Nrep). Although the PDEG values depended on Nrep, we confirmed that both AS and PDEG values were close to zero when samples in the data showed an intermingled nature between the groups in the HSC dendrogram.ConclusionSilhouettes is useful for exploring data with predefined group labels. It would help provide both an objective evaluation of HSC dendrograms and insights into the DE results with regard to the compared groups.Electronic supplementary materialThe online version of this article (10.1186/s12575-018-0067-8) contains supplementary material, which is available to authorized users.

Differential Expression Analysis Using A Model-Based Gene Clustering Algorithm for RNA-Seq Data

“…In a comparison of the performance of each method using different numbers of replicates ( Fig. 1 and Additional le 1), we observed that AUC values tend to increase as the number of replicates increases and this trend is consistent with a previous report [20].…”

“…To date, several methods to enable the analysis of RNA-seq data have been developed, including normalization [5][6][7][8][9][10], various R packages [11][12][13][14][15][16], and graphical user interfaces (GUI) [17][18][19]. Research on more e cient and accurate methods to identify DEGs continues, and new ndings continue to be reported by researchers [20][21][22][23][24][25].…”

Activity-Dependent Remodeling of Synaptic Protein Organization Revealed by High Throughput Analysis of STED Nanoscopy Images

“…The AgglomerativeClustering function from Scikit-Learn with euclidean affinity and Ward linkage was used to perform hierarchical grouping (Pedregosa et al, 2011 ). We calculated the number of groups (synaptic subtypes) that best describe our data based on maximization of the silhouette score, a measure of similarity within a group and dissimilarity between different groups (Rousseeuw, 1987 ; Zhao et al, 2018 ).…”

Commentary: A Systematic Evaluation of Single Cell RNA-Seq Analysis Pipelines

“…However, even the tendency to obtain a large number of DE genes between cell types cannot distinguish these. For example, a bulk RNA-seq dataset exists (Schurch et al, 2016 ) that can produce nearly 70% DE genes (Zhao et al, 2018 ). A common feature of these data sets is a high number of replicates (>40 replicates per group).…”