Journal: Scientific Reports

Article Title: Identifying chromatin features that regulate gene expression distribution

doi: 10.1038/s41598-020-77638-2

Figure Lengend Snippet: Identification of key transcription factors associated with reporter expression levels and noises using multivariate linear regression analysis. ( A ) View of the barcode 29 locus on chromosome 1. Genome coordinate of each barcode was obtained from conducting TRIP experiment. Six examples of normalized ChIP-seq signal profile of transcription factor surrounding the integration site of the barcode 29 were visualized for a domain of 50 kb. ( B ) The heatmap showing the correlation between the enrichment of transcription factors and reporter expression mean and noise. Transcription factors with Spearman's rank-order correlation coefficient of more than 0.2 were selected from over 200 tested transcription factors. ( C , D ) To understand the relationship between the enrichment of TF and the expression levels ( C ) and noise ( D ) of reporter in an integrative way, we selected transcription factors showing a significant correlation in ( B ) to fit multivariate linear regression model. Features with a significant level above the threshold or dashed line (p < 0.05) contributed significantly to the model. ( E ) A Venn diagram listing transcription factors contributing to reporter expression mean (green) and noise (red).

Article Snippet: The analysis of statistical significance was done using standard multivariate regression analysis (using the Matlab command fitlm).

Techniques: Expressing, ChIP-sequencing

Journal: Scientific Reports

Article Title: Identifying chromatin features that regulate gene expression distribution

doi: 10.1038/s41598-020-77638-2

Figure Lengend Snippet: Distance to specific chromatin states influences expression mean and variance ( A ) Nearest distance from the barcode location to each chromatin state segmented by ChromHMM method was calculated. ( B ) The correlation plot between distance to specific chromatin state and reporter expression mean or noise. ( C ) A multivariate linear regression model was used to determine significant chromatin states influencing reporter expression mean and noise. Panels show the student t statistics and p-values of each coefficient in the model for expression mean (top) and noise (bottom). Chromatin states are color-coded.

Article Snippet: The analysis of statistical significance was done using standard multivariate regression analysis (using the Matlab command fitlm).

Techniques: Expressing

Journal: Frontiers in Neural Circuits

Article Title: Cortical Projection From the Premotor or Primary Motor Cortex to the Subthalamic Nucleus in Intact and Parkinsonian Adult Macaque Monkeys: A Pilot Tracing Study

doi: 10.3389/fncir.2020.528993

Figure Lengend Snippet: Multivariable linear regression model (see “Materials and Methods” section). Fit (solid black line) and the 95% confidence bounds (black dashed lines) of the constant (number of BDA-labeled boutons) vs. the other variables (PD symptoms and TH+ loss in SNpc) in the model. The multi variables are “partialed out” but not the constant term. Therefore, the X-abscissa represents the combined independent variables (PD symptoms and TH+ loss) adjusted by the model. The model has an explanatory power (positive slope: y = 9.298* x ; R 2 = 0.821; adjusted R 2 = 0.731; p -value = 0.0117). Three clusters (ovals) appear: (1) the highest TH+ loss and motor impairment (yellow); (2) the lowest loss of TH+ neurons and motor impairment (purple) and (3) the intact animal cluster (gray) without dopaminergic loss nor PD symptoms.

Article Snippet: To assess the relationship between the number of BDA-labeled boutons with the interaction between motor impairment (PD symptoms based on manual dexterity in the modified-Brinkman board task post-ANCE transplantation (see above) and dopaminergic (tyrosine hydroxylase positive (TH+) neurons) loss in the substantia nigra pars compacta (SNpc), we performed a multivariable linear regression model test (MATLAB R2017b, function “fitlm”; Contestabile et al., ).

Techniques: Labeling