TEDDY Metabolomics Reference Analysis

#	Paper	Cohort	Data Modality	Sample (N)	Age / Follow-up	AAB Groups	Key Metabolites	Main Conclusion

Key Findings — Overview

◈

TEDDY NCC1 Dominates the Field

5 of 10 core metabolomics papers draw from the same TEDDY NCC1 nested case-control dataset (~414–418 IA cases, 1:3 matched controls). These are not independent cohorts — results represent convergent analyses from a shared sample pool, not independent replication.

Papers: #3, #4, #5, #6, #7

◎

Pre-seroconversion Window Is the Key Timeframe

All TEDDY NCC1 studies profile plasma at 3-month intervals from birth to first autoantibody appearance. Metabolic signals are detectable as early as 12 months before seroconversion (Balzano-Nogueira 2021), with some signals present at birth (Orešič 2008, DIPP).

Papers: #1, #3, #4, #5, #7

◉

IAA-first vs. GADA-first Have Distinct Metabolic Signatures

TEDDY NCC1 studies consistently distinguish the two seroconversion subtypes. IAA-first is preceded by ↓BCAAs (leucine, isoleucine, valine); GADA-first by ↓proline. This subtype-specificity is a unique advantage of the TEDDY longitudinal design.

Papers: #3, #7

◆

Multi-omics Integration Substantially Improves Prediction

Integrating metabolomics with genetic (SNPs, HLA) and environmental data yields substantially better predictive performance. Webb-Robertson 2021 achieved AUC 0.74; Frohnert 2020 (DAISY) achieved AUC 0.92 for T1D stage progression — metabolomics alone is insufficient.

Papers: #6, #10

▤

Papers #11–16: Background & Methods Context

Papers #11–13 establish clinical/genetic progression benchmarks (JAMA 2013: ~70–84% T1D risk in mAb+ children). Papers #14–16 provide a 2025 TEDDY review, a cross-cohort SM–immune mechanism, and a statistical methods precedent for joint longitudinal-survival modeling.

Papers: #11, #12, #13, #14, #15, #16

#	Paper	Cohort	Primary Modality	Platform / Method	Matrix	N Cases	N Controls	Comparison Groups
1	Dysregulation of lipid and amino acid metabolism precedes islet autoimmunity in children who later progress to T1D Orešič M et al. · J Exp Med · 2008	DIPP	Longitudinal serum metabolomics & lipidomics	GC-MS, LC-MS	Serum	~56 T1D prog.	~56 matched	ND / IA+ / T1D
2	Decreased cord-blood phospholipids in young age-at-onset type 1 diabetes La Torre D et al. · Diabetes · 2013	DiPiS	Cord-blood lipidomics / metabolomics	LC-MS	Cord blood	~56 young T1D	~56 matched	Young-onset T1D vs. controls
3	Longitudinal metabolome-wide signals prior to the appearance of a first islet autoantibody in children participating in the TEDDY study Li Q et al. · Diabetes · 2020	TEDDY NCC1	Longitudinal plasma metabolomics & lipidomics	GC-TOF MS, LC-QTOF MS	Plasma	414 IA cases	~1,242 (1:3)	IAA-first / GADA-first / ND
4	Metabolite-related dietary patterns and the development of islet autoimmunity Johnson RK et al. · Sci Reports · 2019	TEDDY NCC1	Plasma metabolomics & lipidomics + dietary pattern analysis	GC-MS, LC-MS (853 metabolites)	Plasma	352 IA sets	Matched controls	IA / mAb+ / ND
5	Integrative analyses of TEDDY omics data reveal lipid metabolism abnormalities, increased intracellular ROS and heightened inflammation prior to autoimmunity for type 1 diabetes Balzano-Nogueira L et al. · Genome Biology · 2021	TEDDY NCC1	Multi-omics: transcriptomics + metabolomics + dietary biomarkers	RNA-seq, GC-MS/LC-MS, immunoassay	Plasma + PBMC	IA cases (subset)	Matched controls	IA+ vs. ND (5 time points)
6	Prediction of the development of islet autoantibodies through integration of environmental, genetic, and metabolic markers Webb-Robertson B-JM et al. · J Diabetes · 2021	TEDDY NCC1	ML: genetic (SNPs+HLA) + environmental + metabolomics	221-feature ensemble ML	Plasma	IA cases	Matched controls	Persistent IA+ vs. ND
7	Plasma metabolome and circulating vitamins stratified onset age of an initial islet autoantibody and progression to type 1 diabetes: the TEDDY study Li Q et al. · Diabetes · 2021	TEDDY NCC1	Longitudinal lipidome + vitamins + erythrocyte fatty acids	LC-MS + immunoassay + GC	Plasma + RBC	418 IA cases	~1,254 (1:3)	IAA-first / GADA-first / T1D progressors
8	Dynamics of plasma lipidome in progression to islet autoimmunity and type 1 diabetes — DIPP Lamichhane S et al. · Sci Reports · 2018	DIPP	Longitudinal plasma lipidomics	LC-MS	Plasma	40 T1D prog.	40 IA-only + 40 CTR	T1D prog. / IA-only / CTR
9	Metabolomics-related nutrient patterns at seroconversion and risk of progression to type 1 diabetes Johnson RK et al. · Pediatric Diabetes · 2020	DAISY	Metabolomics at seroconversion + dietary nutrient-pattern analysis (RRR)	GC-MS + reduced rank regression	Serum	IA seroconverters	Matched controls	IA → T1D progression
10	Predictive modeling of type 1 diabetes stages using disparate data sources Frohnert BI et al. · Diabetes · 2020	DAISY	Integrated genetic + immunologic + metabolomic + proteomic ML	Multiple platforms, ensemble ML	Serum	NCC subset	Matched (AUC 0.92)	IA onset / T1D stage progression
11	Seroconversion to multiple islet autoantibodies and risk of progression to diabetes in children Ziegler AG et al. · JAMA · 2013	Pooled cohorts	Clinical / immunologic (autoantibody + progression)	Autoantibody assays	Blood	~585 mAb+	Sero-negative	mAb+ vs. single-AAB
12	The influence of T1D genetic susceptibility regions, age, sex, and family history on progression from multiple autoantibodies to T1D: a TEDDY study report Krischer JP et al. · Diabetes · 2017	TEDDY	Clinical / genetic progression analysis	HLA + non-HLA SNP genotyping	Blood	mAb+ children	—	mAb+ → T1D progression
13	A T1D genetic risk score predicts progression of islet autoimmunity and development of T1D in individuals at risk Redondo MJ et al. · Diabetes Care · 2018	TrialNet	Genetic risk score analysis	30-SNP GRS + ImmunoChip	Blood	AAB+ relatives	—	IA → T1D stage
14	Unfolding the mystery of autoimmunity: the Environmental Determinants of Diabetes in the Young (TEDDY) study Rewers M et al. · Diabetes Care · 2025	TEDDY	Review — clinical, omics, and environmental TEDDY outputs	—	—	—	—	Comprehensive TEDDY review
15	Deciphering cross-cohort metabolic signatures of immune responses and their implications for disease pathogenesis Fu J et al. · Mol Syst Biol · 2025	3 non-T1D cohorts	Metabolomics + immune cytokine-response profiling + functional validation	LC-MS + cytokine assays	Plasma	non-T1D subjects	—	Immune-metabolite correlation
16	Joint modeling of longitudinal biomarker and survival outcomes with competing risk in nested case-control studies — TEDDY microbiome dataset Zhao L et al. · Bioinformatics · 2026	TEDDY	Statistical methods — joint longitudinal-survival analysis	Bayesian joint model	Microbiome	—	—	Methods precedent
17	Integration of infant metabolite, genetic, and islet autoimmunity signatures to predict type 1 diabetes by age 6 years Webb-Robertson BM et al. · J Clin Endocrinol Metab · 2022	TEDDY NCC	Integrated ML: infant plasma metabolomics + genetic (HLA/SNPs) + islet autoimmunity	Multi-platform metabolomics + ensemble ML	Plasma	T1D cases	Matched controls	Predict T1D onset by age 6

Key Findings — Data Modality & Cohort

▣

Plasma LC-MS / GC-MS is the Dominant Platform

All TEDDY NCC1 metabolomics studies use plasma sampled at 3-month intervals from birth. Platforms are UC Davis West Coast Metabolomics Center (GC-TOF MS, LC-QTOF MS). This uniformity enables direct metabolite-level comparison across papers #3–#7.

Papers: #3, #4, #5, #6, #7

▤

DIPP Used Serum; DAISY Analyzed at Seroconversion

DIPP studies (Orešič 2008, Lamichhane 2018) used serum from birth cohorts, while DAISY studies analyzed serum at or near seroconversion. These differ from TEDDY's continuous plasma profiling strategy — direct metabolite comparisons across cohorts require caution.

Papers: #1, #8, #9, #10

▦

Sample Sizes Vary Widely Across Cohorts

TEDDY NCC1 is the largest metabolomics resource: 414–418 IA cases with 1:3 matching (~1,200–1,250 controls). DIPP studies used only 40 per group. This disparity means TEDDY findings carry substantially more statistical power for metabolite discovery.

Papers: #3, #7 vs. #8

◈

Multi-omics Studies Add RBC & Transcriptomics Layers

Li Q 2021 (#7) uniquely combined plasma lipidome with erythrocyte membrane fatty acids and circulating vitamins. Balzano-Nogueira 2021 (#5) added PBMC transcriptomics — the only study integrating gene expression with the TEDDY metabolome.

Papers: #5, #7

Data overlap note — TEDDY NCC1 papers (#3, #4, #5, #6, #7): All five papers draw from the same nested case-control dataset. Shared metabolite findings across these papers represent convergent analyses of the same patient sample, not independent replication. True cross-cohort replication exists only when a signal appears in DIPP or DAISY studies as well.

Key Findings — Metabolite Overlap

◈

Phosphatidylcholines Are the Most Replicated Signal

PCs appear in 6 papers across DIPP and TEDDY NCC1, consistently decreased in IA and T1D progressors from birth onward. This is the strongest cross-cohort replicated lipid signal in the T1D metabolomics literature to date.

Papers: #1, #3, #4, #5, #7, #8

◉

Adipic Acid: GEE Analysis Adds Critical Longitudinal Depth

Adipic acid appears in papers #4 and #6 as a risk-associated dicarboxylic acid. The GEE analysis provides the first longitudinal ND → AAB+ → T1D three-stage gradient for this metabolite — the most mechanistically informative report on this signal to date.

Papers: #4, #6 + GEE analysis

◆

Sphingomyelins Link Lipidome to Immune Function

SMs appear in 5 papers and are consistently decreased in progressors. Fu 2025 (#15) provides a mechanistic link, showing SMs negatively regulate monocyte-derived cytokine production across non-T1D cohorts — supporting a functional explanation for the SM signal.

Papers: #4, #5, #7, #8, #15

▣

Triglycerides Show Stage-Dependent Directionality

TG signals are complex: decreased in early infancy (DIPP, #1, #8); unsaturated TGs decreased at seroconversion (#3, #4). The apparent contradiction reflects stage- and subtype-specific dynamics — emphasizing the need for longitudinal, not cross-sectional, designs.

Papers: #1, #3, #4, #7, #8