Decomposing genetic effects of social connectedness and depression on youth behaviors and long-term clinical outcomes

Importance Lacking social connectedness is associated with broad morbidity. Yet the underlying mechanisms remain obscures, especially given its entanglement with depressive symptoms. Clarifying whether social connectedness confers unique risk beyond depression can guide clinical care and social infrastructure interventions. Objective To decompose shared genetic architecture between social connectedness and depressive affect, derive independent polygenic liabilities, and test their differential associations with youth behaviors, digital engagement, and adult clinical outcomes. Design, Setting, and Participants Using summary statisitcs from eight genome-wide association studies (GWAS) on social connectedness related traits, we extracted two orthogonal latent genetic factors: a depression related factor (DEP) and a social disconnection factor (SOC). We performed GWAS-by-subtraction to obtain single nucleotide polymorphisms (SNP) level effect sizes and then constructed factor specific polygenic scores (PGS) in (i) a multi-site longitudinal youth cohort (Adolescent Brain Cognitive Development Study, ABCD, questionnaire set n=11,378; EARS smartphone subsample n=914) and (ii) a large scale electronic health record adult cohort (All of Us Research Program, AoU, n=402,944). We evaluated their relative contributions to youth’s psychopathology, prosocial behaviors, screen time use, and adult’s clinical outcomes. Results In GWAS-by-subtraction, SOC has a GWAS significant loci on TCF4 ; DEP has two loci on TMEM161B and OLFM4 . In youth, a 1-SD increase in SOC PGS was associated with lower prosocial behavior (“being nice to others”; parent-report β=–0.11, SE=0.02, P=7×10 ; youth self-report β=–0.07, SE=0.01, P=4×10) and higher externalizing problems (“argues a lot”; β=0.12, SE=0.02, P=8×10 ¹). In contrast, a 1-SD increase in DEP PGS was more strongly associated with internalizing psychopathology (“too fearful or anxious”; β=0.15, SE=0.02, P=2×10 ¹ ; “unhappy, sad, or depressed”; β=0.14, SE=0.02, P=1×10 ¹) and with fewer high-functioning peers (“friends are athletes”; β=–0.12, SE=0.02, P=3×10 ¹¹; “friends are excellent students”; β=–0.08, SE=0.01, P=4×10 ¹¹); effects replicated directionally in non-European youth. In the EARS subsample, higher SOC and DEP PGS were each associated with greater smartphone screen time. In AoU, SOC PGS was significantly associated with 185 of 712 phecodes (26.0%) and DEP PGS with 167 of 712 (23.5%). Conclusions and Relevance Social connectedness and depressive affect are highly correlated yet genetically separable with distinct behavioral and clinical profiles. Addressing both may be necessary to reduce the health burden associated with the rising reports on loneliness.