Fig. 3: Non-correlating expression markers yield highest diagnostic specificity for schizophrenia patients. | Translational Psychiatry

Fig. 3: Non-correlating expression markers yield highest diagnostic specificity for schizophrenia patients.

From: Dysregulation of a specific immune-related network of genes biologically defines a subset of schizophrenia

Fig. 3

a The correlation matrix depicts present or absent co-regulated gene expression of the top hits in PBMCs derived from human schizophrenia patients and healthy controls. Individual expression levels were correlated using Spearman’s ranked test (group II). + stands for a positive, − for a negative correlation between target expression. Dark blue color indicates correlations that appear in both, SCZ and CTRL subjects; light blue color marks correlations seen only in CTRL subjects, i.e., that were lost in SCZ cases; correlations exclusively appearing in SCZ patients are depicted in red color, color-coding for physiological (light to dark blue) to more pathological (red) relations. All correlation coefficients, P-values and specific n can be found in Supplementary Fig. S9. b Specificity and sensitivity of potential biomarkers RGS1 and CCL4 in the detection of schizophrenia (SCZ) patients. Only cases that showed a target expression lower that 50% of the mean of control (CTRL) cases were counted as detected. By this analysis, information concerning sensitivity and specificity of the targets RGS1 and CCL4 could be gathered. By utilizing RGS1 levels alone, a subgroup of 31% of the SCZ cases could be detected, with a false positively detecting 12% of CTRL subjects. CCL4 analyzed in that manner identified a subgroup of 39% of SCZ patients with a specificity of 95% for detection of SCZ patients. A combination of both biomarkers led to a specificity of 97% and a sensitivity of 27%. CTRL control subjects, SCZ schizophrenia patients

Back to article page