SUPPLEMENTARY TABLE S4.

Listing of some of the most novel or impactful findings from the HERITAGE Family Study from 1995 to 2021.

Novel or Impactful Findings	Trait(s)	Details	Citation(s)
Heterogeneity of training responses
Large interindividual differences in phenotype responses to endurance training	V̇O2max	V̇O2max responses ranged from no gain to >1 L·min−1 (−5% to +48%).	(43,44,74)
	IVGTT traits	The proportion of subjects who had no change or decreases in IVGTT traits ranged from 42% to 55%.
No correlation between baseline V̇O2max and its changes with training	V̇O2max	The correlation between baseline and delta V̇O2max was 0.08, with baseline levels after adjustments explaining only 2% of the variance in delta V̇O2max.	(33,43)
No aggregation of training responsiveness in individuals or subgroups	V̇O2max	There were low, average, and high V̇O2max responses across ages, both sexes, both ethnic groups, and levels of initial V̇O2max.	(43)
	V̇O2max, % fat, AVF, insulin, HDL-C, small LDL, GlycA	Almost half of the cohort had at least one high response and one low response across seven cardiometabolic traits. There is a high degree of trait specificity in training responsiveness.	(283)
Some unfavorable responses to endurance training	Fasting insulin; HDL-C; TG; SBP	Prevalence of potentially “adverse” response in HERITAGE ranged from 6% to 9% for the four traits.	(75)
Familial aggregation of baseline phenotype levels
Within-subject variability	V̇O2max	Within-subject SD from measures repeated days and weeks apart range from 108 to 137 mL O2·min−1 with CV of 4.1% to 5.0%.	(20,27,33,34), plus unpublished data
Significant familial resemblance for intrinsic V̇O2max	V̇O2max	2.7 times more variance between families than within families. Heritability of intrinsic V̇O2max was 51%, with significant maternal heritability (29%).	(45)
Significant familial aggregation for intrinsic submaximal V̇O2	V˙O2 at 50 W, 60%, 80%	Maximal heritability estimates ranged from 48% to 70% and 29% to 48% for maternal heritability.	(48)
Significant familial aggregation for intrinsic submaximal Q and SV	SV and Q at 50 W and 60% V̇O2max	Maximal heritability estimates of 46% for SV and Q at 60% V̇O2max and 41% and 42% at 50 W	(58)
Significant familial aggregation for visceral fat	AVF	Heritability of 47% to 48%, independent of total body fat	(32)
Higher heritability of resting BP in Black subjects	Resting SBP and DBP	Heritabilities of SBP and DBP were 68% and 56% in Black subjects vs 43% and 24% in White subjects.	(56)
Significant familial aggregation for muscle enzyme activities	CK, PHOS, HK, PFK, GAPDH, LPL, CPT, HADH, CS, COX	Strong familial aggregation for activities of baseline muscle enzymes related to PCr, glycolytic, and oxidative metabolism.	(138,143)
Significant familial aggregation for markers of oxidative stress	LDL-ox, C50-AAPH, TBARS, glutathione	Heritability for oxidative stress traits ranged from 31% to 44%.	(127)
Familial aggregation of training response
Significant familial aggregation for V̇O2max training response	V̇O2max	Heritability was 47%, with maternal inheritance accounting for 28%.	(41)
Significant familial aggregation for submaximal V̇O2 training response	V̇O2 at 50 W, 60%, 80%	Heritability values ranged from 23% to 57% for the training response of submaximal measures of V̇O2.	(48)
Significant familial aggregation of submaximal exercise BP and HR training responses	HR50, HR60%, SBP50	Heritabilities for HR50, HR60%, and SBP50 responses were 34%, 29%, and 22%. Heritability for DBP traits and all HR and BP traits were lower in Black subjects.	(71)
Significant familial aggregation for submaximal Q & SV training responses	SV & Q at 50 W and 60% V̇O2max	Heritabilities ranged from 24% to 38% for SV and Q at 50 W and 60% V̇O2max.	(58)
Significant familial aggregation of interindividual variation in plasma lipid responsiveness to training	TC, TG, LDL-C, apoB, HDL-C, HDL2-C, HDL3-C, apoA-I	Heritability ranged from 25% to 38% for lipid response traits. Exceptions were for heritability levels near 60% for changes in apoB in Blacks and HDL2-C in Whites and a lack of heritability for change in LDL-C in Black subjects.	(98)
Significant familial aggregation for training response of markers of oxidative stress	LDL-ox, C50-AAPH, TBARS, glutathione	Heritability for oxidative stress training response traits ranged from 35% to 84%.	(127)
Significant familial aggregation for training response of muscle enzyme activities	CK, PHOS, HK, PFK, GAPDH, CPT, HADH, CS, COX	Strong familial aggregation was found for training response of muscle enzymes related to PCr, glycolytic, and oxidative metabolism.	(138,143)
Ethnic and sex differences in responses to training
Training responses of submaxmial exercise measures of hemodynamic traits differed by sex and ethnic groups	HR, SBP, DBP at 50 W	Submaximal HR, SBP, and DBP decreased with training, with greater reductions in women compared with men and in Black and older subjects compared with White and younger subjects.	(17), pp. 10–116
	a-V̇O2 diff, SV, Q, and V̇O2 at 50 W	Black men did not increase a-V̇O2 diff at 50 W. Thus, on average, Black men had greater increases in SV50 and smaller decreases in Q50 compared with White men to achieve similar VO250.	(61), MSSE, pp. 99–106
Significant sex interactions for insulin sensitivity response to training	Si	Insulin sensitivity increased by 10% in the total cohort, with the increase larger in men than women (16% vs 5%).	(74)
Significant ethnic and sex differences in lipid, lipoprotein, and lipase activity responses to training	ApoA-I, HDL2-C	ApoA-I increased more in women than men, in Black than White subjects, and in offspring than in parents. Black subjects experienced greater increases in HDL2-C compared with White subjects.	(84)
	LPL activity	LPL activity increased in all subgroups except Black men.	(97)
Response of other traits to training
Little influence of endurance training on markers of oxidative stress	LDL-ox, C50-AAPH, TBARS, antioxidants, and aminothiols	Only erythrocyte resistance to hemolysis significantly changed with training, which interacted with smoking status (smokers did not experience beneficial effects of training on erythrocytes), and was significant in women only.	(127)
No change in RMR with training.	RMR via indirect calorimetry	Sample size (N = 77) was larger than previous studies. There was no change in RMR at 24 or 72 h after training.	(153)
Favorable changes in clinically relevant lipoprotein subfractions in response to training	NMR-based lipoprotein subfractions	Large HDL-P and LDL-P increased, whereas small LDL-P and all VLDL subfractions and VLDL-P size decreased with training. These findings were not captured with traditional lipid profiling (i.e., TG and LDL-C did not change in total sample).	(93)
Genome-wide linkage studies of baseline and response phenotypes
Identification of QTLs for baseline V̇O2max and for V̇O2max response	V̇O2max	QTLs on 4q, 8q, 11p, and 14q were reported for baseline V̇O2max. QTLs on 1p, 2p, 4q, 6p, and 11p were identified for change in V̇O2max.	(210)
Dense mapping of 4 QTLs identified strong candidate genes for submaximal exercise capacity and hemodynamic responses to training	SV50, HR50	Titin, kinesin family member 5B, cAMP responsive element binding protein 1, MIPEP, and SGCG genes were identified as strong candidates for changes in submaximal SV and HR and submaximal exercise capacity.	(211,231,233)
First GWAS of exercise response traits
GWAS of V̇O2max response to endurance training	V̇O2max	39 SNPs were associated at P < 1.5 × 10−4, with a panel of 21 SNPs accounting for 49% of the variance in V̇O2max trainability.	(234)
GWAS of submaximal HR response to training	HR50	40 SNPs associated at P < 9.9 × 10–5, with top hit 8 × 10−7. Nine SNPs accounted for the genetic variance of the submaximal exercise HR response to training.	(235)
Molecular signatures of exercise response derived from integrative analyses of genomic and transcriptomic profiles
Muscle gene expression and SNP signatures predict V̇O2max response to endurance training	V̇O2max	Genome-wide baseline muscle gene expression and validation identified a 29-RNA signature that predicted changes in V̇O2max. Candidate genes from this predictor and the literature led to a 11-SNP signature that explained 23% of the variance in V̇O2max trainability.	(265)
Combined genome-wide and transcriptome-wide analysis identifies SNPs associated with TG response to training	TG	GWAS identified 4 SNPs accounting for the genetic variance of TG response, whereas molecular signature based on the baseline expression of 11 genes predicted 27% of TG changes in response to training. An 8-SNP score comprising 4 SNPs each from transcriptomics and GWAS was the strongest predictor of TG training response.	(236)
GWAS and transcriptional signature of insulin sensitivity response to training	Si	Integrative analysis of functional genomic and transcriptomic profiles identified combined variation in genes linked to cholinergic, calcium, and chemokine signaling associated with Si training response. MEF2A transcription factor was the most significant candidate driving the transcriptional signature associated with ∆Si, strengthening the relevance of calcium signaling in exercise training-mediated Si response.	(238)
Proteomic signatures of V̇O2max and its trainability
Plasma protein signature of cardiorespiratory fitness level	V̇O2max	Elastic net regression identified 115 proteins highly correlated (r2 = 0.80) with measured V̇O2max, which was replicated in the validation data set, with an r2 of 0.71 (Fig. 9).	(271)
Plasma proteins associated with intrinsic V̇O2max and its trainability	V̇O2max	147 proteins were associated with baseline V̇O2max, whereas 102 baseline proteins were associated with changes in V̇O2max, with minimal overlap (only 5 proteins) between protein sets. A baseline 56-protein signature improved prediction of V̇O2max response (AUC 0.84) compared with a model of only clinical variables (AUC 0.62).	(272)
Bioinformatics explorations of intrinsic V̇O2max and its trainability
Integrative pathway analysis of V̇O2max response to training GWAS	V̇O2max	Using GWAS results followed by candidate gene prioritization and pathway analysis, pathways related to calcium signaling, energy sensing and partitioning, mitochondrial biogenesis, angiogenesis, immune functions, and regulation of autophagy and apoptosis were identified as key mechanisms through which the physiological responses of V̇O2max to training are mediated.	(273)
Genetics and biology underlying intrinsic V̇O2max	V̇O2max	A bioinformatics pipeline applied to V̇O2max data in the sedentary state suggests 4 loci related to cardiovascular physiology (ATE1, CASQ2, NOTO, and SGCG), four loci related to hematopoiesis (PICALM, SSB, CASQ2, and CA9), 4 loci related to skeletal muscle function (SGCG, DMRT2, ADARB1, and CASQ2), and 8 loci related to metabolism (ATE1, PICALM, RAB11FIP5, GBA2, SGCG, PRADC1, ARL6IP5, and CASQ2) as candidates for human variability in cardiorespiratory fitness among sedentary adults.	(275)
Metabolomic biomarker of training responsiveness
DMGV is a biomarker of metabolic response to endurance training	Fasting glucose, insulin, and lipids	Baseline levels of plasma DMGV associated with lack of improvement in HDL traits. DMGV levels decreased with training and were positively correlated with several lipid, glucose, and insulin traits	(96)