Covariate sensitivity in donor-level reanalysis of human islet transcriptomes in type 2 diabetes

Background Inflammatory and metabolic abnormalities in type 2 diabetes (T2D) islets are established, but interpretation of a particular donor cohort also depends on covariate conditioning and cell capture. This reanalysis assessed which aspects of the reported associations were sensitive to these choices. Methods GSE50244 whole-islet expression (51 non-diabetic [ND] and 11 T2D donors) and GSE221156 single-cell data (17 ND and 17 T2D donors) were analyzed separately at donor level. Age/sex and, additionally, BMI-adjusted expression models were compared using coefficients and uncertainty diagnostics. Captured-cell composition was examined with centered log-ratio models, technical covariates, alternative pseudocounts, and 200 library-level downsampling draws. Results Bulk disease coefficients were correlated across the two adjusted models (Spearman rho = 0.807), although the median unmoderated standard-error ratio after adding BMI was 1.190. Thus, the larger retained limma discovery count did not correspond to a general improvement in ordinary least squares precision. Captured beta-cell proportions averaged 43.72% in ND and 31.04% in T2D. No individual cell type met BH FDR <0.05 in the BMI-adjusted composition model, while the joint composition test depended on technical adjustment and pseudocount choice. The beta-cell CLR coefficient remained negative in all downsampling draws, but the global test crossed P < 0.05 in 113 of 200 draws. Retained expression summaries were consistent with established inflammatory and beta-cell metabolic patterns. Conclusion The additional contribution is a quantitative assessment of how model conditioning and cell sampling affect the interpretation of these cohorts. It separates directional consistency from precision and significance as well as relative captured-cell shifts from evidence of tissue cell loss. BMI-conditioned estimates are descriptive associations, not adiposity-independent causal effects. These cohort-specific observations provide a basis for future independent replication and mechanistic investigation.

Authors

Institutions

Publication Details

Journal
Frontiers in Endocrinology
Published
2026-11-19
DOI
https://doi.org/10.3389/fendo.2026.1994040
Primary Topic
Pancreatic function and diabetes
Type
article
Field-Weighted Citation Impact
0.00
Controls
|||
ALL TIME
JAN
FEB
MAR
APR
MAY
JUN
JUL
AUG
SEP
OCT
NOV
article

Covariate sensitivity in donor-level reanalysis of human islet transcriptomes in type 2 diabetes

Dan Chen, Wenqian Huang
Frontiers in Endocrinology
Pancreatic function and diabetes
article

Covariate sensitivity in donor-level reanalysis of human islet transcriptomes in type 2 diabetes

Dan Chen, Wenqian Huang
article en

Abstract

Background Inflammatory and metabolic abnormalities in type 2 diabetes (T2D) islets are established, but interpretation of a particular donor cohort also depends on covariate conditioning and cell capture. This reanalysis assessed which aspects of the reported associations were sensitive to these choices. Methods GSE50244 whole-islet expression (51 non-diabetic [ND] and 11 T2D donors) and GSE221156 single-cell data (17 ND and 17 T2D donors) were analyzed separately at donor level. Age/sex and, additionally, BMI-adjusted expression models were compared using coefficients and uncertainty diagnostics. Captured-cell composition was examined with centered log-ratio models, technical covariates, alternative pseudocounts, and 200 library-level downsampling draws. Results Bulk disease coefficients were correlated across the two adjusted models (Spearman rho = 0.807), although the median unmoderated standard-error ratio after adding BMI was 1.190. Thus, the larger retained limma discovery count did not correspond to a general improvement in ordinary least squares precision. Captured beta-cell proportions averaged 43.72% in ND and 31.04% in T2D. No individual cell type met BH FDR <0.05 in the BMI-adjusted composition model, while the joint composition test depended on technical adjustment and pseudocount choice. The beta-cell CLR coefficient remained negative in all downsampling draws, but the global test crossed P < 0.05 in 113 of 200 draws. Retained expression summaries were consistent with established inflammatory and beta-cell metabolic patterns. Conclusion The additional contribution is a quantitative assessment of how model conditioning and cell sampling affect the interpretation of these cohorts. It separates directional consistency from precision and significance as well as relative captured-cell shifts from evidence of tissue cell loss. BMI-conditioned estimates are descriptive associations, not adiposity-independent causal effects. These cohort-specific observations provide a basis for future independent replication and mechanistic investigation.

Frontiers in EndocrinologyVol. 17
Ocean University of China (CN)
Pancreatic function and diabetes
AI Navigator

Ask Laika to Summarize, Analyze, and Connect papers live on the map.

Summarize Papers & Methodologies

Extract key findings, datasets, and comparative methods across publications.

Benchmark Rankings & Visual Analytics

Rank top research institutions, authors, funders, topics, and journals by Field-Weighted Citation Impact (FWCI) and paper volume with instant charts.

Connect Distant Disciplines

Bridge topological clusters on the map to find hidden collaborative intersections.

Laika access coming soon