BACKGROUND:The potent immunogenicity of mRNA vaccination raised concerns about its impact on HIV-1 viral load (VL) and reservoir size in people with HIV, which could drive vaccine hesitancy. As part of a prospective clinical trial, we investigated whether SARS-CoV-2 mRNA vaccination induced changes in VL and reservoir size after a 1-year three-dose regimen in people with HIV on antiretroviral therapy (ART). METHODS:We collected blood samples from 37 people with HIV with varying degrees of immune reconstitution before vaccination and after receiving their second and third doses of vaccine. Anti-SARS-CoV-2-Spike antibodies and Spike-specific CD4+T-cells were analysed as indicators of vaccine-induced immunogenicity. The Intact Proviral DNA Assay was used to quantify the intact and defective HIV-1 reservoir. RESULTS:Vaccination with three doses of SARS-CoV-2 mRNA vaccine over the course of 1 year did not result in plasma HIV-1 RNA rebound nor significantly alter the HIV-1 reservoir size regardless of immune status. The median intact HIV-1 reservoir size was 53 (IQR 34-193) copies/million CD4+T-cells at baseline and 86 (34-268) and 138 (33-334) 3 months after the second and third dose, respectively (p > 0.38). We found no correlation between changes in VL or HIV-1 reservoir size and the anti-SARS-CoV-2-Spike levels, Spike-specific CD4+T-cells, CD4+/CD8+ ratio, nadir or CD4+T-cell count. However, the intact reservoir size negatively correlated with CD4+T-cell count, CD4+/CD8+ ratio and nadir CD4+T-cell count. CONCLUSIONS:Our findings show that three doses of SARS-CoV-2 mRNA vaccine did not affect the VL or HIV reservoir size regardless of immune status, supporting their safety in people with HIV. Notably, a higher CD4+T-cell count correlated with a reduced intact HIV-1 reservoir size.

Alterations in the gut bacterial communities are well described in people with HIV-1 (PWH). However, less is known about the gut viral communities (virome). Herein, we performed an exploratory study of the gut DNA virome in PWH receiving antiretroviral therapy (PWH-ART). Virome profiling was performed using two complementary viral taxa identification pipelines, MetaPhlAn4 and Phanta, providing high-resolution taxonomic profiles at the species level. Consistent patterns of the gut DNA virome were obtained with both pipelines, though Phanta identified a greater number of viral taxa at the species level. Most gut DNA viruses were unclassified species, and bacteriophages constituted the majority of identifiable viral populations. Compared to HIV-1 negative controls (n = 11), PWH‑ART (n = 13) exhibited compositional differences in their gut DNA virome at finer taxonomic levels (species and genus), including higher within-group variations and multiple differentially abundant viral taxa. Gender influenced the virome composition, independent of HIV-1 status and age. Furthermore, correlations were identified between multiple unclassified viral species and clinical variables, including gender, CD4+ T cell counts, and ART duration. Further studies are warranted to validate these findings and to elucidate whether the observed compositional differences in the gut DNA virome are mediated by HIV-1 infection itself, by ART, or by both.

BACKGROUND:Although COVID-19 is no longer a public health emergency, it remains the most prevalent circulating infectious-like-illness in Europe. Whether immunocompromising conditions (ICCs) still carry increased mortality risk during the Omicron era is unclear. METHODS:We conducted a cohort study across EuCARE sites in 8 countries among adults admitted to hospital with COVID-19 between 2020-2023. ICCs and COVID-19 pneumonia at hospitalization were defined using clinical information and ICD-10 codes. Logistic regression and counterfactual mediation analysis was used to compare 28-day in-hospital mortality risk associated with ICCs using COVID-19 pneumonia and vaccination at hospital entry as intermediates. Proportion of the total effect of ICCs mediated and the controlled direct effects (CDEs) were calculated. We also formally tested for interaction between SARS CoV-2 variants and ICCs for mortality risk. FINDINGS:42,488 individuals were included, of which 1,675 (3.9%) had an ICC. 55% were male, median (IQR) age was 67 (52, 79) years. Overall, 4,344 (10.2%) individuals died in hospital. ICCs were associated with increased mortality, OR = 1.49 (1.25, 1.79) with no evidence for an attenuation during the Omicron phase (p-interaction=0.60). Mediation analyses showed that the total effect of ICCs was mediated by vaccination but only weakly by pneumonia. With Omicron, the excess mortality associated with ICC was higher under the scenario that everyone in the cohort was to develop COVID-19 pneumonia [CDE =1.22 (0.09, 1.65)]. INTERPRETATION:ICC remains a significant risk factor for in-hospital death, even during the Omicron era, particularly if the infection led to the development of pneumonia.

Background:HIV-1 sub-subtype A6 is predominant in Eastern Europe and was associated with increased risk of treatment failure with the long-acting cabotegravir plus rilpivirine regimen. In this study, we aimed to evaluate the in vitro susceptibility and the genetic barrier to resistance to INSTI in recombinant viruses harboring clinically derived A6 integrase coding regions. Methods:We generated 23 NL4-3 strain-based recombinant viruses harboring clinically derived integrase coding region. We measured their susceptibility to second-generation INSTIs dolutegravir, bictegravir, and cabotegravir in a TZM-bl cell-based phenotypic assay. The genetic barrier to resistance was evaluated by exposing MT-2 cell cultures infected with 4 A6 integrase recombinant viruses, as well as the NL4-3 and HXB2 subtype B reference strains. Results:All 23 recombinant viruses generated with clinically derived A6 integrase displayed full susceptibility to dolutegravir, bictegravir, and cabotegravir, showing median (interquartile range) fold-change values of 1.2 (0.9-1.5), 1.1 (0.7-1.5), and 0.9 (0.6-1.1), respectively. Of 4 A6 viruses assessed for their genetic barrier to resistance in vitro, only 1 showed emerging integrase mutations E138K or Q148R at subinhibitory concentrations of dolutegravir or cabotegravir, respectively. Conclusions:These data suggest that sub-subtype A6 integrase has full susceptibility and largely maintains a high genetic barrier to resistance to second-generation integrase strand transfer inhibitors.

OBJECTIVE:The objective of the study was to establish whether HIV-1 sub-subtype A6 (HIV-1A6) is a risk factor for virological failure in people with HIV (PWH) treated with the high genetic barrier integrase strand transfer inhibitors (INSTIs) dolutegravir (DTG) or bictegravir (BIC). METHODS:The virological outcome of first-line DTG or BIC-containing antiretroviral therapy (ART) was assessed in 261 people with HIV-1A6 (PWH-1A6) and 1042 people with HIV-1B (PWH-1B) starting treatment between January 2014 and May 2025 with follow-up for at least one year in the EuResist Integrated Database. RESULTS:Most PWH-1A6 were recent migrants from Ukraine. The event rates per 100 person-years follow-up were higher in PWH-1A6 for low-level viremia (LLV, 2.71 vs. 1.71, P = 0.049) and virological failure with more than 1000 HIV RNA copies/mL (4.12 vs. 2.03, P < 0.001) than in PWH-1B. At the end of follow-up, 226/261 (86.6%) PWH-1A6 and 936/1042 (89.8%) PWH-1B had viral load below 50 copies/ml (P = 0.132). INSTI DRMs were observed in 5/219 (2.3%) available integrase sequences of PWH-1A6, including two cases detected at virological failure with more than 1000 HIV RNA copies/ml and two cases in virologically suppressed PWH-1A6. In total, 12/261 PWH-1A6 discontinued DTG- or BIC-containing ART, including three individuals who were not virologically suppressed at discontinuation. CONCLUSION:In PWH treated with DTG or BIC-containing first-line ART, LLV and virological failure with more than 1000 HIV RNA copies/ml were observed more frequently in PWH-1A6 than in PWH-1B. However, such events rarely resulted in INSTI resistance or discontinuation of INSTI-containing ART.

The gut microbiome is altered during HIV-1 infection and contributes to immune dysfunction and inflammation in people living with HIV (PLWH), these changes may persist despite effective antiretroviral therapy (ART). We explored the associations between the fecal gut microbiome and blood HIV-1 reservoir size in PLWH (n = 30) on long-term ART. The intact proviral DNA assay (IPDA) and shotgun metagenomic sequencing were performed to identify microbial species and metabolic pathways associated with the size of the HIV-1 reservoir. PLWH with a smaller intact reservoir exhibited lower evenness compared to individuals with a larger intact reservoir. We found that Phocaeicola plebeius and Lachnospira sp000437735 were significantly enriched in individuals with a smaller intact reservoir and lower intact-to-total proviral ratio, respectively. We observed a negative association between Faecalibacterium prausnitzii and a positive association of Prevotella copri, with the intact proviral reservoir size. Additionally, the metabolic pathways of glycolysis and branched-chain amino acid biosynthesis were enriched in individuals with larger reservoir. HIV reservoir size in blood is associated with gut microbiome evenness, specific metabolic pathways and microbial signatures, including Lachnospira, Prevotella, and Faecalibacterium. Our findings underscore the potential role of the gut microbiome in viral persistence, raising the possibility that modulating microbial composition could influence the HIV reservoir.

ABSTRACT Despite advances in treatment, HIV-1 infection continues to remain a major global health challenge, prompting ongoing efforts to understand the mechanisms that enable natural viral suppression and immune control. Elite controllers (ECs), a rare subset of PLWH individuals, naturally suppress HIV-1 replication without antiretroviral therapy, highlighting the importance of host-related factors in viral control. Understanding the mechanisms underlying this unique phenotype is crucial for developing novel therapeutic strategies. Previous studies from our group identified certain EC-specific metabolites, called dipeptides (DPs), and investigated their antiviral properties. We hypothesize that these dipeptides may potentially affect epithelial barrier integrity by modulating the expression of tight junction proteins, which in turn influences the mucosal barrier function, a key factor in HIV-1 pathogenesis. Therefore, in this study we investigated the impact of ten EC-specific DPs on tight junction (TJ) gene and protein expression in epithelial models derived from the female reproductive and gastrointestinal tracts, where we observed enhanced expression of different TJ genes ( CLDN1, CLDN3, CLDN4, CLDN7, CLDN14, TJP1, TJP2, OCLN ) and proteins (CLDN1, CLDN7, and CLDN14), suggesting the potential influence of these dipeptides on epithelial barrier function. Furthermore, we also examined different proteomic profiles between dipeptide (WG)-treated HeLa CD4 + CCR5 + cells compared with the untreated ones, and observed significantly reduced abundance of pro-inflammatory proteins, such as RELB Proto-Oncogene (RELB), TNF-α-induced protein 1 (TNFAIP1), TNF receptor superfamily member 1A (TNFRSF1A), and IL-32, in dipeptide-treated cells; and increased expression of proteins associated with tissue homeostasis (SMAD family member 5 [SMAD5]), cellular proliferation (transforming growth factor β receptor 3 [TGFBR3]), and epithelial integrity, like CD81. Interestingly, KEGG analysis revealed possible attenuation of NF-κB, MAPK, TNF, and JAK-STAT signaling pathways, along with the enrichment of mTOR and PI3K-AKT pathways in treated HeLa CD4 + CCR5 + cells. Overall, this study investigated the potential interplay between tight junction proteins and key signaling pathways involved in maintaining epithelial barrier integrity and modulating immune activation, potentially contributing to both HIV-1 control and to the chronic inflammation associated with infection.

Motivation:The emergence of multidrug class resistance (MDR) in Human Immunodeficiency Virus (HIV) is a rare but significant challenge in antiretroviral therapy (ART). MDR, which may arise from prolonged drug exposure, treatment failures, or transmission of resistant strains, accelerates disease progression and poses particular challenges in resource-limited settings with restricted access to resistance testing and advanced therapies. Early prediction of future MDR development is important to inform therapeutic decisions and mitigate its occurrence. Results:In this study, we employ various machine learning classifiers to predict future resistance to all four major antiretroviral drug classes using features extracted from clinical HIV sequence data. We systematically explore several variations of the problem that differ in the pre-existing resistance level and the temporal gap between sample collection and observed MDR occurrence. Our models show the ability to predict multidrug class resistance even in the most challenging variations, albeit at a reduced accuracy. Feature importance analysis reveals that our models primarily utilize known drug resistance mutations for easier classification tasks, but rely on new mutations for the difficult task of distinguishing four class drug resistance from three class drug resistance. Availability and implementation:All analysis was performed using the Euresist Integrated DataBase (EIDB). Researchers wishing to reproduce, validate or extend these findings can request access to the latest EIDB release via the Euresist Network.

HIV evades immune detection through rapid mutation of its surface proteins, yet essential steps in viral entry, such as CD4 and co-receptor engagement, remain highly conserved. While therapies like Lenacapavir represent major advances, the emergence of resistant strains highlights the urgent need for adaptable, rapid-response antivirals. This challenge extends beyond HIV, demanding scalable design strategies for diverse viral threats. Here, we demonstrate that AI-driven design can address this need by generating cyclic peptide binders targeting a previously unexploited interface on the HIV-1 fusion protein gp41. Using only sequence information, without prior structural or binding site data, we designed and experimentally validated a single candidate. This inhibitor potently blocked infection by two HIV-1 strains in cell-based assays with no detectable cytotoxicity. Affinity analysis with SPR confirms the interaction with gp41 as designed. Our findings illustrate how AI-guided peptide design, coupled with rapid in-vitro validation, can accelerate early-stage therapeutic discovery and enable timely intervention against emerging viral threats.

The persistence of HIV-1 latency reservoirs in CD4+ T cells is a significant obstacle for curing HIV-1. Shock-and-kill strategies, which aim to reactivate latent HIV-1 followed by cytotoxic clearance, have shown limited success in vivo due to insufficient efficacy of latency reversal agents (LRAs) and off-target effects. Natural killer (NK) cells, with their ability to mediate cytotoxicity independent of antigen specificity, offer a promising avenue for enhancing the shock-and-kill approach. Previously, we observed that pan-caspase inhibitors induce NK cells to secrete an LRA in vitro. Here, we aimed to identify this LRA using a targeted proteomic approach. We identified lymphotoxin-α (LTα) as the key LRA secreted by NK cells following pan-caspase inhibitor treatment. LTα was shown to significantly induce HIV-1 LTR promoter activity, a hallmark of viral reactivation. Neutralization of LTα effectively abolished the observed LRA activity, confirming its central role. Moreover, cytokine-primed but not resting human primary NK cells exhibited LRA activity that could be neutralized with LTα neutralizing antibodies. Finally, pan-caspase inhibitor treatment did not decrease the ability of the cytokine-primed NK cells to kill target cells. These findings demonstrate that cytokine-primed NK cells, through LTα secretion, can effectively reactivate latent HIV-1 following pan-caspase inhibitor treatment, without compromising NK cell cytotoxicity. This highlights a potential enhancement strategy utilizing NK cells for shock-and-kill approaches in HIV-1 cure research.

HIV-1 infection cannot be cured due to the presence of HIV-1 latently infected cells. These cells do not produce the virus, but they can resume virus production at any time in the absence of antiretroviral therapy. Therefore, people living with HIV (PLWH) need to take lifelong therapy. Strategies have been coined to eradicate the viral reservoir by reactivating HIV-1 latently infected cells and subsequently killing them. Various latency reversing agents (LRAs) that can reactivate HIV-1 in vitro and ex vivo have been identified. The most potent LRAs also strongly activate T cells and therefore cannot be applied in vivo. Many LRAs that reactivate HIV in the absence of general T cell activation have been identified and have been tested in clinical trials. Although some LRAs could reduce the reservoir size in clinical trials, so far, they have failed to eradicate the reservoir. More recently, immune modulators have been applied in PLWH, and the first results seem to indicate that these may reduce the reservoir and possibly improve immunological control after therapy interruption. Potentially, combinations of LRAs and immune modulators could reduce the reservoir size, and in the future, immunological control may enable PLWH to live without developing HIV-related disease in the absence of therapy.

HIV evades the immune system through rapid mutation of its surface proteins, particularly the envelope glycoprotein. However, the core mechanism of viral entry, CD4 binding and co-receptor engagement remains conserved. While therapies such as Lenacapavir represent important advances, the continued emergence of resistant strains will demand new and more adaptable treatment strategies. This challenge is not unique to HIV; future pandemics will likely present similar pressures, highlighting the need for drug design methods that are not only effective but also fast and scalable. Recent advances in protein structure prediction have transformed the landscape of therapeutic design, enabling the accurate modelling of target structures from sequence alone and now facilitating the development of novel therapeutics without prior structural data. EvoBind leverages these advances to rapidly generate cyclic peptide binders in a single design round, using only the amino acid sequence of a target protein. Cyclic peptides offer several advantages over traditional linear protein molecules, including increased stability, while their small size improves oral bioavailability and enables access to challenging binding sites. Here, we demonstrate the use of EvoBind to generate cyclic peptide binders against the HIV envelope protein gp41, which is essential for viral-host membrane fusion. Cell-based assays confirm potent inhibition of two different HIV-1 strains with no detectable toxicity. The combination of artificial intelligence-guided design and streamlined experimental validation can significantly accelerate therapeutic development, reduce costs, and provide timely solutions to the challenges posed by viral evolution and emerging global health threats. ### Competing Interest Statement P.B. is a shareholder in Cyclic Tx.

AIMS:Lenacapavir (LEN) is a first-in-class HIV-1 capsid inhibitor. We investigated the natural occurrence of LEN drug resistance mutations (DRM) in HIV-1 sequences from LEN treatment-naïve individuals. METHODS:We analysed all available HIV-1 capsid sequences (n = 21 646) from individuals never exposed to LEN from the Los Alamos National Laboratory database. RESULTS:LEN DRMs were identified at all LEN DRM-associated positions, including the highly resistant M66I variant, which was found in 9 sequences across multiple subtypes. In total, 56 sequences (0.26%) contained LEN DRMs. The most prevalent mutation was Q67H (n = 22), and this mutation was more common in CRF01_AE (17/3053, 0.56%). CONCLUSIONS:All known LEN DRMs can be found in viral populations without drug pressure. This finding may reflect immunological pressure on capsid epitopes containing DRMs. Our findings are of potential clinical relevance for LEN use in both therapeutic and prophylactic applications.

Background There is limited data on whether SARS-CoV-2 infections will result in increased long-term use of general healthcare, potentially impacting healthcare systems and management. Exploring this, we investigated the healthcare use of individuals with a SARS-CoV-2 infection in 2020 over a period of two years, using comprehensive medical records.Methods We followed a cohort of 365,354 individuals in Stockholm, Sweden, who had been tested with SARS-CoV-2 serology in 2020, for healthcare use during 2021/22. SARS-CoV-2 seropositive and seronegative individuals were matched 1:1 on age, sex, 2019 healthcare use, and date of last serology, and compared on healthcare use during 2021/22 using registry linkages. Seropositive individuals were stratified on hospitalization for COVID-19 in 2020. Individuals were compared for total healthcare use, measured as incidence rate rations (IRR), and healthcare type usage-or-not per month, measured as a difference-in-differences regression.Results There were 272,918 seronegative and 73,814 seropositive subjects. Incidence rate ratios (IRRs) for primary healthcare use were 1.0, 1.16, and 0.98, for all, only hospitalized, and only non-hospitalized, seropositive individuals respectively. For outpatient specialist care IRRs were 0.96, 1.31, and 0.93. For inpatient care IRRs were 0.98, 1.19, and 0.95. Healthcare type usage-or-not per month showed no substantial differences, ranging from 0.01 to -0.01 in deviation. Increased healthcare use during follow-up was restricted to the seropositive individuals hospitalized for COVID-19 in 2020.Conclusion There was no increase in healthcare use in the overall population from SARS-CoV-2 infections during 2020, suggesting there is no apparent need to adapt healthcare systems at scale for the COVID-19 aftermath.

Background:HIV viremia has been considered a cardiovascular disease (CVD) risk factor, but many studies have had insufficient data on potential confounders. We explored the association between viremia and CVD after adjusting for established risk factors and analyzed whether consideration of viremia would improve CVD prediction. Methods:Adults from RESPOND were followed from the first date with available data until the first of rigorously defined CVD, loss to follow-up, death, or administrative censoring. We first analyzed the associations between 6 measures of viremia (time-updated, before antiretroviral therapy [ART], viremia category, and measures of cumulative viremia) and CVD after adjusting for the variables in the D:A:D CVD score (age, sex/gender, smoking, family history, diabetes, recent abacavir, CD4 count, blood pressure, cholesterol, high-density lipoprotein, cumulative use of stavudine, didanosine, indinavir, lopinavir, and darunavir). We subsequently compared predictive performance with and without viremia in 5-fold internal cross-validation. Results:A total of 547 events were observed in 17 497 persons (median follow-up, 6.8 years). Although some viremia variables were associated with CVD in univariable analyses, there were no statistically significant associations after adjusting for potential confounders, neither for measures of current viral load, pre-ART viral load, highest viremia category during ART, nor cumulative viremia (modeled both as total cumulative viremia, cumulative viremia during ART, and recent cumulative viremia). Consistently, none of the viremia variables improved prediction capacity. Conclusions:In this large international cohort, HIV viremia was not associated with CVD when adjusting for established risk factors. Our results did not show viremia to be predictive of CVD among people with HIV.

BACKGROUND:Limited evidence exists on how bacterial and viral coinfections have developed since the SARS-CoV-2 Omicron variant emerged. We investigated whether community-onset coinfections in adult patients hospitalized with COVID-19 differed during the wild type, Alpha, Delta, and Omicron periods and whether such coinfections were associated with an increased risk of mortality. METHODS:We conducted a multinational cohort study including COVID-19 hospitalizations until 30 April 2023 in 5 European countries. The outcome was bacterial and viral coinfections based on 5 test modalities. Variant periods were compared with regard to occurrences of coinfections and risk ratios for coinfections (Omicron vs pre-Omicron), as well as association with in-hospital mortality (Omicron vs pre-Omicron). RESULTS:A total of 29 564 cases were included: 12 601 wild type, 5256 Alpha, 2433 Delta, and 9274 Omicron. The coinfection rate was 2.6% (327/12 601) for wild type, 2.0% (105/5256) for Alpha, 3.2% (77/2433) for Delta, and 7.9% (737/9274) for Omicron. Omicron had a significantly increased risk ratio of coinfection when compared with preceding variants (1.88 [95% CI, 1.53-2.32], P < .001). These results were consistent across several subgroup analyses. An increased occurrence (19% [232/1246] vs 11% [3042/28 318]) and adjusted risk (1.69 [95% CI, 1.49-1.91], P < .001) of in-hospital mortality were observed in patients with a verified coinfection as compared with patients without a coinfection. CONCLUSIONS:Bacterial and viral coinfections were more prevalent during the Omicron period as compared with preceding variants. Such coinfections were associated with an increased risk of in-hospital mortality, calling for sustained monitoring and clinical vigilance.

Background Despite effective antiretroviral treatment (ART), HIV infection is associated with immune dysfunction and inflammation. Metformin has shown beneficial immunological and anti-inflammatory effects, including in people with HIV (PWH). We studied the potential association between metformin treatment and immune reconstitution in PWH.Methods We conducted a retrospective cohort study set in Stockholm, Sweden. PWH with T2DM who initiated metformin treatment after at least 2 years on effective ART (exposed individuals) and metformin-na & iuml;ve PWH (controls) were matched in a 1:1 ratio based on age, sex, baseline immune status, and duration of ART. Outcomes included mean values of CD4 cell counts and CD4/CD8 ratios from 1.5 years to 3.5 years after compared with 2 years before the exposed individual started metformin treatment (index date).Results Among 1332 PWH, 43 metformin-exposed individuals (median age, 48 years; 11 years since start of ART) with T2DM and 43 nondiabetic controls (median age, 47 years; 11 years since start of ART) were included in the matched analyses. The median (interquartile range) change in CD4 T-cell count was 35 (-21 to 125) cells/mu L among exposed individuals and 48 (-18 to 100) cells/mu L among controls (P = .96). The corresponding numbers were 0.10 (0.03 to 0.20) and 0.08 (0.02-0.16) for CD4/CD8 ratio (P = .18). No differences were observed in subgroup analyses of PWH with low CD4 T-cell counts and CD4/CD8 ratios.Conclusions No significant differences in immune reconstitution were observed between metformin-treated individuals and matched controls over the 2-year follow-up period. We observed no significant differences in immune reconstitution (measured by CD4 T cell count and CD4/CD8 ratio) between metformin treated individuals and matched controls over a 2-year follow-up period.

Predicting the outcome of antiretroviral therapies (ART) for HIV-1 is a pressing clinical challenge, especially when the ART includes drugs with limited effectiveness data. This scarcity of data can arise either due to the introduction of a new drug to the market or due to limited use in clinical settings, resulting in clinical dataset with highly unbalanced therapy representation. To tackle this issue, we introduce a novel joint fusion model, which combines features from a Fully Connected (FC) Neural Network and a Graph Neural Network (GNN) in a multi-modality fashion. Our model uses both tabular data about genetic sequences and a knowledge base derived from Stanford drug-resistance mutation tables, which serve as benchmark references for deducing in-vivo treatment efficacy based on the viral genetic sequence. By leveraging this knowledge base structured as a graph, the GNN component enables our model to adapt to imbalanced data distributions and account for Out-of-Distribution (OoD) drugs. We evaluated these models' robustness against OoD drugs in the test set. Our comprehensive analysis demonstrates that the proposed model consistently outperforms the FC model. These results underscore the advantage of integrating Stanford scores in the model, thereby enhancing its generalizability and robustness, but also extending its utility in contributing in more informed clinical decisions with limited data availability. The source code is available at https://github.com/federicosiciliano/graph-ood-hiv

The human blood proteome provides a holistic readout of health states through the assessment of thousands of circulating proteins. In this study, we present a pan-disease resource to enable the study of diverse disease phenotypes within a harmonized proteomics dataset. By profiling protein concentrations across 59 diseases and healthy cohorts, we identified proteins associated with age, sex, and body mass index, as well as disease-specific signatures. This study highlights shared and distinct protein patterns across conditions, demonstrating the power of a unified proteomics approach to uncover biological insights. The dataset, covering 8262 individuals and up to 5416 proteins, serves as an online resource for exploring disease-specific protein profiles and advancing precision medicine research.