Fig. 3
From: Large-scale labeling and assessment of sex bias in publicly available expression data
Y chromosome loss is prevalent and variable across cell lines. a Cell line sex label switching. The sex of the donor cell line is on the left and the imputed sex is on the right. Samples are divided into female (blue) and male (orange). Additionally for donor cell sex labels, we include samples with unknown sex (gray) and samples with metadata mapping to more than one cell line (green). b Distribution of average study sex scores (e.g. P(male) for a sample) for male cell lines shows a bimodal pattern, indicating that many of these cell lines appear “female-like” (see Additional file 2: Table S7 for a list). c Profiles of Y chromosome loss across male cell lines with more than five studies. Each cell line is a row. From left to right, the first panel shows the fraction of male (orange), female (blue), and unlabeled (gray) samples, the second shows the distribution of study average sample sex scores, and the third shows the cell line specific distributions of Y chromosome gene copy number (CNV) from the Cancer Cell Line Encyclopedia [31]. Overall, this demonstrates the highly variable and cell line-specific nature of Y chromosome loss