Janice (Yibing) Jiang

1. Transcriptome characterization of human immune gene regulation in human AD

• Use Python for ST analysis of 230 human brain samples across three cortical regions (CR, PFC, VC) to investigate C1Q-mediated immune regulation in human aging and Alzheimer’s disease.
• Performed EDA and demographic stratification to identify confounding effects on C1Q gene expression; removed confoundings via transcriptome-wide residualization using linear modeling.
• Identified C1QA, C1QB, and C1QC co-regulators; executed gene ontology enrichment analysis of C1Q-centered networks, revealing a shift from adaptive immune regulation in controls to complement activation, and neutrophil-associated inflammation in AD.

2. Spatial transcriptional characteristics of C1q-centered immune genes in AD mouse models.

• Performed differential gene expression analysis on different age mouse models; identified a significant, slide‐wide downregulation in 6-month samples (both WT and AD).
• Adapted scLVM’s framework for spatial transcriptomics (SGLBC), selecting housekeeping genes as guide genes, computing latent “batch” factors via PCA on residuals, and regressing out batch effects per gene.
• Implemented α-shape–based landmark extraction and Procrustes rigid alignment to register spot coordinates from all slides into a common reference frame, correcting for slide-level spatial misalignment.

• Collected and curated a cohort of 16,670 singleton deliveries from 2011 to 2025 at Beijing 306 Hospital, excluding multiple gestations, and integrating key demographic, weight, and complication variables.
• Classified participants by pre-pregnancy BMI and GWG categories per CDC standards, and extracted and one-hot encoded the top 10 pregnancy complications (e.g., GDM, pre-eclampsia, PROM) from unstructured text.
• Performed exploratory analyses of complication incidence, trimester-specific and total weight-gain trajectories, and maternal age trends in Chinese singleton pregnancies from 2011 to 2024.
• Generated stratified baseline summaries by BMI group and weight-gain category, and conducted subgroup analyses across BMI, weight-gain, age cohort, parity, and height intervals—visualizing results with forest plots.
• Applied four complementary approaches to derive optimal gestational weight-gain rates: interquartile subgroup bounds, incidence-minimizing intervals, adjusted odds-ratio modeling, and restricted cubic spline regression.