Skip to main content

The association between single-nucleotide polymorphisms within type 1 interferon pathway genes and human immunodeficiency virus type 1 viral load in antiretroviral-naïve participants



Human genetic contribution to HIV progression remains inadequately explained. The type 1 interferon (IFN) pathway is important for host control of HIV and variation in type 1 IFN genes may contribute to disease progression. This study assessed the impact of variations at the gene and pathway level of type 1 IFN on HIV-1 viral load (VL).


Two cohorts of antiretroviral (ART) naïve participants living with HIV (PLWH) with either early (START) or advanced infection (FIRST) were analysed separately. Type 1 IFN genes (n = 17) and receptor subunits (IFNAR1, IFNAR2) were examined for both cumulated type 1 IFN pathway analysis and individual gene analysis. SKAT-O was applied to detect associations between the genotype and HIV-1 study entry viral load (log10 transformed) as a proxy for set point VL; P-values were corrected using Bonferroni (P < 0.0025).


The analyses among those with early infection included 2429 individuals from five continents. The median study entry HIV VL was 14,623 (IQR 3460–45100) copies/mL. Across 673 SNPs within 19 type 1 IFN genes, no significant association with study entry VL was detected. Conversely, examining individual genes in START showed a borderline significant association between IFNW1, and study entry VL (P = 0.0025). This significance remained after separate adjustments for age, CD4+ T-cell count, CD4+/CD8+ T-cell ratio and recent infection. When controlling for population structure using linear mixed effects models (LME), in addition to principal components used in the main model, this was no longer significant (p = 0.0244). In subgroup analyses stratified by geographical region, the association between IFNW1 and study entry VL was only observed among African participants, although, the association was not significant when controlling for population structure using LME. Of the 17 SNPs within the IFNW1 region, only rs79876898 (A > G) was associated with study entry VL (p = 0.0020, beta = 0.32; G associated with higher study entry VL than A) in single SNP association analyses. The findings were not reproduced in FIRST participants.


Across 19 type 1 IFN genes, only IFNW1 was associated with HIV-1 study entry VL in a cohort of ART-naïve individuals in early stages of their infection, however, this was no longer significant in sensitivity analyses that controlled for population structures using LME.


Disease progression among antiretroviral treatment (ART) naïve people living with HIV (PLWH) differs substantially [1], with both viral and host genetic factors (and their interaction) known to play an important role [2,3,4]. Human leukocyte antigen (HLA)-type is the most established host genetic factor associated with HIV progression [1, 3]. Additionally, variation within genes encoding chemokines and chemokine receptors, including CCR5D32, are known to be important for HIV susceptibility, primarily for HIV viral load and disease progression [1, 5]. However, common genetic variation, including HLA and CCR5 SNPs, together with participant demographic variables are cumulatively estimated to explain only ~ 25% of VL variability [3, 6, 7]. From investigations of heritability in HIV-1 disease progression, the contribution of host genetics´ on HIV-1 viral load variability is estimated to be ~ 30% [8], of which about 5% is from loci beyond the well characterized variants in the major histocompatibility complex and CCR5 regions of the genome. As such, a gap of knowledge remains. Other host genetic factors affecting HIV progression have been theorised to consist of either a combination of many small effect size variants or involve more complex interactions of variants across a genetic pathway [1, 7, 9]. Thus, here we explore the use of other approaches to investigate the role of host genetics in HIV pathogenesis beyond the traditional genome wide association studies (GWAS) [1].

Pathway analysis can assess genetic variants that may act in a functionally dependent manner and discover complex associations with an impact on the variability in disease progression [10]. In contrast to the multiple comparisons made in GWAS´, where variants are analysed individually, in pathway analysis groups of variants are analysed simultaneously, which increases the power to detect associations [10]. As such, a pathway analysis can exploit the advantages of analysing all variants (e.g., SNPs) contained within a genetic pathway. Knowledge of genetic pathways allows one to select SNPs from the distinct genes within the pathway to investigate their collective association with the outcome of interest. For HIV-1 disease progression, one such potential pathway is the type 1 interferon (IFN) pathway.

Type 1 interferons are important cytokines in the host´s immune system that play a role in combating viral infections. In SARS-CoV-2 infection, host genetic variation of type 1 IFN genes have been shown to impact the course of the infection [11,12,13]. For HIV, the type 1 IFN pathway is of particular interest [14,15,16]. In HIV, these cytokines not only mediate an early response as one of the major parts of the innate immune system, they also induce the transition of the innate immune system to the adaptive immune system [17, 18]. Further, type 1 IFNs induce important viral restriction factors in the large number of IFN stimulated genes. As such, the type 1 interferon pathway is a portion of the interferon pathway.

To identify IFN-induced HIV restriction factors M. OhAinle et al. [9] performed a CRISPR gene neutralization of these genes in CD4+ T cells. They found the inhibition of HIV-1 replication by IFNs were caused by combined actions of only a few IFN stimulated genes. These are known as HIV restriction factors, e.g., MxB, TRIM5alpha, IFITM1 and Tetherin.

Prior GWAS’ in HIV-1 infection including both European and multi-ethnic populations have not identified SNP level associations with HIV viral load (HIV-VL; a proxy of disease progression) in type 1 IFN genes [7, 19]. However, it may be that multiple SNPs across the individual type 1 IFN genes or the entire pathway impact HIV replication.

The stable level of viral load, defined as set point VL (spVL), which appears during the asymptomatic phase after acute HIV-1 infection, is a common prognostic measure for HIV infection severity, which can predict both infectiousness and rate of progression to disease [2]. Although, the level of HIV RNA before initiation of ART has been shown to differ significantly between groups based on sex, risk, age at- as well as year of seroconversion, and presentation during acute infection [20]. Using data from the Strategic Timing of Antiretroviral Treatment (START) trial [21], this study included a gene and pathway level analysis of the type 1 IFN pathway to assess whether accumulated genetic variation across individual genes or the entire pathway affects HIV replication in individuals with early infection. We then used data from the Flexible Initial Retrovirus Suppressive Therapies (FIRST) cohort [22] to validate the findings from the START trial in individuals with more advanced infection.


Participant population

The early infection cohort are participants from the START (NCT00867048) [23] trial, which enrolled HIV-1 positive individuals between 2009 and 2013 across a global network of clinical sites. At study entry, individuals were ART-naïve, aged > 18 years, and with CD4+ T-cell count > 500 cells/mm2. Relevant clinical data and biological material for research (among those who consented to specimen collection) were collected at the participant recruiting site. The FIRST cohort [22] was used for validation and included participants with more advanced HIV-1 infection, as there was no CD4 count requirement. The FIRST trial enrolled HIV positive participants from the US between 1999 and 2002, who were ART-naïve at study entry. The analyses in both cohorts were restricted to those who gave consent for genetic analysis.

Outcome in START and FIRST

Viral load at study entry was used as the primary outcome measure in both the START and FIRST cohorts [21, 24].

Genes included and SNP selection

The seventeen type 1 IFN pathway genes and their two receptors (IFNAR1, IFNAR2) located on either chromosome 9 or 21 were selected through an in-house software using pathway information (Table 1), such as gene members from the type 1 IFN pathway sourced from the Reactome database [25, 26]. Gene loci information was retrieved from Ensembl database [27, 28] through the BioMart R package [29]. Any SNPs overlapping the genes in the type 1 IFN pathway as well as being within a window of 2000 nucleotides upstream and downstream from each gene were included.

Table 1 Type 1 IFN genes and SNP counts

Study cohort genotype data

Genotypic data from the START and FIRST trials has been described previously [21, 22]. Briefly, human DNA from study participants´ blood samples were genotyped using a custom Affymetrix Axiom SNP-array (including 770,558 probes), which was enriched using immune dysfunction related markers. Genome Reference Consortium Human Build 37 (Hg19) in Ensembl gene database [27, 28] was used for gene annotation.

Quality control (QC) of genetic data

Participants were excluded if they had any of the following: sex mismatch, autosome SNP call rate less than 96%, duplicates, cryptic relatedness estimated by pairwise identity-by-descent (IBD) (pi-hat at least 0.90).

SNP Quality Control: SNPs with any of the following were kept for the GWAS: (a) recommended by Axiom Analysis Suite (Thermo Fisher Scientific), (b) >  = 90% reproducibility from the internal control Ref103, (c) For multiallelic sites, the SNP record with the higher quality score was kept and the remaining SNPs occurring at the same position were removed.

SNP Imputation and post imputation QC

SNP imputation to the 1000 Genomes phase 3 [30] (genome build: GRCh37) was performed on raw genotyping data with the genipe pipeline [31] using PLINK (v2.00a3LM) [32] and SHAPEIT (version 2.5) [33] for phasing and IMPUTE2 (version 2.3.2) [34] for imputation. Imputed SNPs with a confidence score INFO <  = 0.8 and duplicates were removed. The threshold for minor allele frequency (MAF) was decreased from a limit of MAF > 5% in the previous GWAS [21] to MAF > 1% for the inclusion of rare SNPs. Remaining post-imputation QC was performed following the previous publication [21]. Briefly, SNPs fulfilling any of the following were excluded: (a) genotype missing rate > 10%, or (b) Hardy–Weinberg equilibrium p-value < 1 × 10−6. After SNP imputation, subjects fulfilling any of the following: (a) > 10% missing SNPs or (b) outside of expected heterozygosity (i.e., those with F values outside of 3 standard deviations above or below the mean) were also excluded.

Calculation of PCAs and ancestry estimation

EIGENSTRAT [35] was used for principal components analysis (PCA) and the top 4 eigenvectors were included as covariates in the analysis to control for population stratification (Additional file 1). PCA was performed independently in each cohort based on SNPs passing QC and having MAF > 0.05 in line with the previous GWAS studies [21, 24]. For the START cohort only, continental-level ancestry was estimated using ADMIXTURE ( and included 1000 Genomes samples (phase 3; n = 2504) as a reference. Inclusion criteria required that the SNPs were common to both the HIV cohort and 1000 genomes reference samples, array genotyped, non-ambiguous, biallelic, autosomal, and not present in high LD regions (as defined in or in HLA region chr6: 28477797–33448354 (hg19/GRCh37). PLINK (v1.9 and v2.0) was used for SNP and sample filtering (–maf 0.01, –geno 0.1, –hwe 1e−6, –mind 0.1) and pairwise-pruning (window size of 50 variants, step size of 5 variants, and r2 threshold of 0.2), leaving 159 406 SNPs for ancestry estimation. ADMIXTURE was run with K ancestral populations set to five, for consistency with the number of 1000 Genomes superpopulations.

Statistical analyses

Preparation and performing SKAT-O analysis

The sequence kernel association test (SKAT) is a powerful method for gene level analysis of SNP array data and pathway analysis [10, 36]. SKAT-O [37,38,39,40] is an optimised variant, and is a bidirectional approach, which combines features of dispersion tests (i.e., allow for different directions of effect of rare variants on the outcome) and burden tests (i.e., assume effects of all rare variants having the same direction on outcome). Detection of associations through SKAT-O was between the genotypes collapsed into gene regions (here SNPs within type 1 IFN genes or receptors) or cumulatively in the pathway analysis of all 19 genes, and the phenotype; HIV-1 study entry VL.

To make use of this method, R scripts were implemented around SKAT-O where data formatting was performed to optimise the input data prior to using SKAT-O to test for associations with HIV-1 study entry VL. Briefly, START and FIRST cohort data were reformatted into three data sources; (1) Affymetrix SNP array probe identities remapped to Reference SNP cluster IDs (rsID) using chromosome and position (for imputed SNPs that did not have Affymetrix array probe id, their original rsID was used instead); (2) Participants´ allele counts for each SNP id; (3) Participants´ phenotype information and clinical data. These datasets were used to generate SNP sets lying within gene co-ordinates given by entrez ID [41] and the additional window size. Since Affymetrix array data used probe ID to identify SNP calls, data source (1) allowed the retrieval of Affymetrix probe IDs using rsID together with chromosome and position information. Data source (2) then allowed the access to retrieve participants´ genotype and subsequently, testing the cohort data against the phenotype of interest with corresponding covariates using data source (3). The scripts were written with R software (V. 4.2.0) in addition to using R SKAT-O package [42, 43] and named as GeneAnalysis_SKATO.R with GeneAnalysis_SKATO_Helper.R and checkDependencies.R as supporting scripts. Instructions for obtaining the docker image as well as access to the scripts are available at

Entrez ID [41] for type 1 IFNs and IFNAR1, IFNAR2 were provided as input data to the R scripts together with the upstream and downstream window size; covariates (gender and the first four principal component vectors), and the outcome (participants´ study entry HIV-1 viral load, log10 transformed); herein referred to as the main model. To test multiple genes (i.e., a gene set), Entrez IDs were grouped to be treated as one test set. SKAT-O runs were configured using the recommended default settings.

SNPs data and selection of SNPs included for SKAT-O analysis

Restrictions were placed to only include type 1 IFN genes and interferon-α receptors (IFNAR1, IFNAR2) in an initial investigation of type 1 IFNs. When SNPs were cross checked with gene loci in Ensembl [27, 28], a total of 673 SNPs were retrieved for the type 1 IFN genes, of which 625 (92.8%) were imputed SNPs (Table 1). Each SNP overlapped one of the 19 type 1 IFN pathway genes within an upstream/downstream window of 2000 base pairs of each gene’s loci (Table 2). The distribution of all 673 SNPs in IFN Genes are in Additional file 2.

Table 2 Demographics of participants at study entry

REGENIE method

SNP filtering, LD pruning, and REGENIE Step 1 were performed based on the full cohort (n = 2429). Step 1 of REGENIE [44] was run on biallelic, directly genotyped autosomal SNPs that were filtered with PLINK (–maf 0.01, –geno 0.1, –mind 0.1, –hwe 1e−15) and pruned for LD (window size of 1000 variants, step size of 100 variants, and r2 threshold of 0.9). Gender and the first 4 PCs (standardized to mean = 0 and standard deviation = 1) were included as covariates and the phenotype was log10 transformed. This step used a block size of 1000. SKATO analysis (Step 2) was run separately for the full cohort and two African groups that were defined based on geography (n = 339) or genetics (n = 525). The SKATO analysis used the same covariates and phenotype as in Step 1 and focused on 673 SNP variants in 19 genes of interest.

Sensitivity analysis

To assess the robustness of associations, sensitivity analyses were performed with additional covariates potentially related to viral load. These consisted of age; baseline CD4+ T-cell count and CD4+ / CD8+ T-cell ratio, and whether participants had a recent HIV infection (i.e., within six months before enrolment) were included separately in SKAT-O analysis. Recent HIV infection was calculated based on participants´ self-reporting and a multi-assay algorithm used on baseline samples to verify these participants as having a recent infection [45]. Further, since the START cohort is geographically diverse, we performed an additional sensitivity analysis using the REGENIE method, which used LME to control for population structures. Here, the subgroup analysis was based on participants´ geography (Europe, U.S., Latin America, Africa) and by gender (Table 2). Australia was not included due to low number of participants (n = 91). We also conducted a subgroup analysis of persons of African descent as defined by their ADMIXTURE fraction cut off (at least 0.7). Finally, we performed additional subgroup sensitivity analyses using the REGENIE method on the African subgroup defined by continent of recruitment and by ADMIXTURE fraction cutoff of at least 0.7.

Performing individual SNP association

Individual SNP level association was performed for any genes found to be significantly associated with HIV-1 viral load, using a linear model through PLINK (v2.00a3LM). When performing individual SNP association analysis, the same input parameters used in SKAT-O analysis with the addition of age were provided as covariates. Variance was standardized for age as a covariate parameter.

Multiple testing correction

Bonferroni correction was used to limit the family-wise error rate. A total of 20 SKAT-O analyses were performed: 19 individual analyses of the sets of SNPs within the type 1 IFN and receptor genes and one cumulative analysis including all type 1 IFN pathway SNPs (gene set). Using Bonferroni’s correction, the adjusted significance level was P < 0.0025 in correspondence to a total of 20 comparisons.


Using the FIRST cohort for validation, SKAT-O was applied for the type 1 IFN pathway analysis or in any type 1 IFN gene showing a significant p-value in START after adjustment for covariates.


START participants

A total of 2440 START participants were eligible for inclusion in the analysis. However, since post imputation QC identified 11 participants with a heterozygosity coefficient (F value) outside ± 3 standard deviations from the mean, the total number of START participants included in this analysis decreased to 2429 (Table 2). Most participants were white (n = 1398, 58%), male (n = 1940, 80%), and enrolled in Europe (47%). The median age at study entry was 37 years (IQR 25 to 45). The median time since HIV diagnosis was 1.1 years (IQR 0.4–3.0) with a median viral load of 14,623 copies/mL (IQR 3460–45100) at enrolment. As per the enrolment criteria, all individuals had a CD4+ cell count > 500 cells/µL (median 651, IQR 585–760). The prevalence of hepatitis B and hepatitis C co-infections was low, with, respectively, only 2.3% and 4.5% of START participants being HIV-1/hepatitis co-infected.

SNPs within IFNW1 region were significantly associated with study entry VL in START cohort

Cumulated SKAT-O analysis of type 1 IFN gene and receptor SNPs (i.e., all 19 genes treated as one set) showed no significant association (p = 0.15) with study entry VL (Table 3). In the individual gene analysis, SKAT-O detected a borderline significant association between higher levels of study entry VL and IFNW1 (p = 0.0025). No other type 1 IFN gene was found to be significantly associated with study entry VL. Results were consistent in sensitivity analysis where additional covariates, including age, CD4+ T-cell count, CD4+/CD8+ T-cell ratio and recent infection were included (Table 3). The sensitivity analysis using the REGENIE method as an additional control for population structure were largely consistent with the main results, however the p-value for IFNW1 was no longer significant after adjustment for multiple testing (p = 0.0244).

Table 3 SKAT-O P-values for type 1 IFN gene associations with the START cohort study entry VL

Two subgroup analyses were performed; one stratified by the four main geographical locations: Europe (N = 1135); Africa (N = 339); Latin America (N = 413); and the U.S. (N = 451), and the other by gender (Table 4). In subgroup analysis by geography, the association between IFNW1 and HIV-1 study entry VL was only replicated in the African subgroup (p = 0.002). No significant association was detected between IFNW1 and study entry VL in the other geographical subgroups. In the sensitivity analyses of the African subgroup, where we controlled for population structure using linear mixed effects models in addition to the PCs, the p-value for the association with VL increased to p = 0.0614. In the subgroup where persons of African descent were identified via Admixture, the p-value for the association with VL was slightly lower (p = 0.0348). In the analysis by gender, no significant outcome p-value was detected. However, it is worth noting 47% (N = 232) of the females included in the START study were enrolled in Africa compared with only 5% (N = 107) of the males.

Table 4 Subgroup analyses of IFNW1 association with study entry VL using SKAT-O

Investigation and assessment of IFNW1 SNPs in START

To further investigate the SNPs within the IFNW1 region, individual SNPs were extracted from imputation data. Seventeen SNPs were retrieved (Table 1). MAFs of the 17 IFNW1 SNPs are shown in Table 5. One SNP, rs10964859, was in the three prime untranslated region (3´-UTR) of IFNW1 gene. The remaining 16 SNPs overlapped either 2000 base pairs upstream or downstream from the IFNW1 gene. IFNW1 is in chromosome 9, located at positions 21,140,631 to 21,141,831 in the GRCh37 genome build.

Table 5 Alternate allele frequencies of the 17 IFNW1 SNPs in START and GnomAD in percentages

Linkage disequilibrium (LD) of IFNW1 SNPs in START

We investigated Linkage Disequilibrium (LD) of the 17 IFNW1 SNPs in the START cohort. Heatmap shows most of the SNPs were independent with a cluster of SNPs (rs10964861, rs10964860, rs10611694 and rs10757189) in LD (Fig. 1, Additional file 3).

Fig. 1
figure 1

Heatmap illustrating IFNW1 SNPs linkage disequilibrium (LD) in the START cohort. Legend indicates the strength with red representing high LD

Individual SNP association assessment with study entry VL in the START cohort by use of the same covariates (PC1, PC2, PC3, PC4 and gender) revealed no significant association among any of SNPs analysed, except for rs79876898, an A to G substitution (beta = 0.32, p = 0.002), which was found to be associated with higher viral load (Table 6). One of 17 SNPs, rs1895673 was included in the previously reported GWAS [21], while the remaining 16 SNPs were imputed and hence only examined here. The outcome of rs1895673 in the GWAS (p = 0.048, beta = − 0.059) was similar to the result for the SNP (p = 0.037, beta = − 0.062) in this study.

Table 6 Association analysis of 17 IFNW1 SNPs with study entry VL in the START cohort

Validation in FIRST cohort

For the FIRST cohort, 544 participants had available genotyping data. Three participants were excluded due to an F value outside ± 3 standard deviations from the mean. This resulted in a final cohort of 541 participants. Similar to the START cohort, the majority of the participants in the FIRST cohort were male (N = 431, 80%) and the median age was 38 years old (IQR 32–44 years old). In contrast, self-reported ethnicities of this cohort were primarily black (N = 309, 57%) with the next largest group being white (N = 142, 26%). All participants were enrolled from the US. Time since HIV diagnosis was unavailable. The FIRST trial consisted of ART naïve individuals with advanced disease, and this is reflected in study entry characteristics, particularly median HIV VL, which was higher in the FIRST cohort when compared to the START cohort (126,298 vs 14,623 copies/mL). The prevalence of hepatitis B (5%) and hepatitis C co-infections (17%) at study entry in FIRST was also higher in comparison to START (2% and 5%, respectively).

No significant association was detected between IFNW1 and study entry VL (p = 0.1665) in FIRST participants when tested under the same conditions as START participants (Table 7). Also, sensitivity analysis after adjustment for age and CD4+ T-cell count at baseline, respectively, did not lead to a significant outcome of IFNW1 SKAT-O. As we did not observe associations using the main model, we did not conduct subgroup analysis of the FIRST participants. Likewise, we did not investigate associations of single IFNW1 SNPs with HIV VL in the FIRST cohort.

Table 7 SKAT-O p-values of IFNW1 association with the study entry VL in the FIRST cohort


In this study, we conducted a pathway analysis with SKAT-O using phenotype and genotype data from a diverse cohort of ART naïve PLWH to access associations between type 1 IFN pathway genotypes and HIV-1 study entry VL. While the pathway analysis did not show a significant association in analysis of individual type 1 IFN genes, a borderline significant association between IFNW1 and study entry VL (p = 0.0025) was detected that was stable in sensitivity analysis when additional covariates potentially related to HIV-VL were added. However, in the sensitivity analysis utilising linear mixed effects models to control for population structure, the p-value increased above the multiple testing corrected threshold in the full START cohort. Subgroup analysis indicated this association could be driven by participants from Africa, although again, the signal was lower in subgroup analyses using linear mixed effects models to control for population structure. Therefore, this signal warrants further investigation in a cohort from this region. In analysis of IFNW1 single SNP associations, only rs79876898 was significantly associated with study entry VL (higher if G than A).

This study builds on a previous GWAS in START [21] by incorporating a gene and pathway level analysis of the type 1 IFN pathway. No significant association with HIV study entry VL was observed in analysis of the type 1 IFN pathway. However, the association with study entry VL of one type 1 IFN gene, IFNW1, which encodes the interferon omega-1 protein, indicates that the combination of the 17 IFNW1 SNPs may influence HIV VL. Thus, genetic variation in IFNW1 may be of importance for downstream type 1 IFN signalling and thereby impact HIV-1 replication. SKAT-O assesses bidirectional associations, but the methodology does not provide any information on the direction of the potential effect of IFNW1 on viral replication (upregulate/downregulate). To gain some knowledge on this we subsequently investigated the individual SNP level associations in the IFNW1 gene. Here, one SNP, rs79876898 was found to be associated with higher viral load. This SNP was not examined in the previous GWAS [21], as that study did not impute SNPs. We additionally searched for this SNP in other GWAS publications [3, 46] that focused on the association of variants with HIV viral load. Chromosome 6 was a recurring finding in terms of association with HIV viral load. However, rs79876898 (located in chromosome 9) was not identified in any of the studies. Furthermore, inspection of dbSNP [47] databases returned records for rs79876898; but no publications were linked to the SNP. ClinVar [48], a database that records association of SNPs to clinically relevant diseases also did not yield any records. Hence, the identification of rs79876898 to be associated to higher viral load appears to be novel, but the lack of prior associations and the potential confounding by population structures mean that this association should be interpreted cautiously.

IFNW1 is known to have antiviral, anti-proliferation, and antitumor effects [49]. In one in vitro study IFNW1 was shown to be a potent inhibitor of HIV replication, and compared with IFNA2, HIV-1 protein synthesis was more sensitive to IFNW1 in this study [50].

Three of the IFNW1 SNPs had prior citations in PubMed [51]; rs10757189, rs10511694, and rs10964859 [52,53,54,55] but not related to HIV. rs10964859 C > G is an imputed SNP located in 3´-UTR of IFNW1 and having MAFs of 35%, 26% and 19% in European, African and Latino/admixed American populations, respectively [56]. In studies investigating cancer patients, rs10964859 has been noted as a potential regulatory variant, which may impact gene expression due to loss of miRNA binding [52,53,54]. The two additional SNPs, rs10757189 [55], rs10511694 [53, 55] have previously been reported to be associated with outcomes related to cancer; melanoma (rs10511694) and colorectal cancer (rs10757189, rs10511694). We did not find any prior citations for the remaining SNPs. By investigation of LD in the START cohort, we found rs10757189 and rs10511694 were in LD (r2 = 0.913). rs79876898, which was significantly associated with study entry VL in START, was not in LD with any of the three IFNW1 SNPs, that had been published. Future studies may lead to improved knowledge of whether these SNPs impact type 1 IFN gene expression or type 1 IFN proteins in vitro or in vivo.

In subgroup analysis the association of IFNW1 with study entry VL appeared to be driven by participants of African descent. Different gender compositions between geographical subgroups may affect the results of SKAT-O subgroup analysis, since VL is known to be affected by gender [57]. Further, different HIV-1 subtypes between geographical subgroups may have affected the VL levels [58]. The finding related to a possible association between variation in IFNW1 and study entry VL in START [23] was not validated in the FIRST cohort [22]. In addition to having a much lower sample size, the lack of consistency of findings between these cohort may be explained by the composition of the two cohorts. FIRST exclusively enrolled in the U.S, compared to START where only 20% were from the U.S. Additionally, individuals in FIRST were more advanced in their HIV infection than START participants (with substantially lower CD4+ T-cell count and higher HIV-VL at study entry). The sensitivity analysis adding baseline CD4+ T-cell count as a covariate did not alter the outcome of IFNW1 SKAT-O in the FIRST cohort. However, the stages of progression in HIV-1 infection seen in START [23] and FIRST [22] might be controlled by different type 1 IFN genes. Additionally, the proportions of participants coinfected with Hepatitis B or C differed between START and FIRST participants. Notably, the proportion of FIRST participants co-infected with HCV was 17%. However, in a review and meta-analysis, HIV/HCV co-infection was not found to lead to a significantly higher HIV VL [59].

The potential association of IFNW1 with HIV-1 VL was also not detected in a study by McLaren and colleagues [7], which used SKAT-O to identify associations between all human genes and HIV progression. The cohort investigated by McLaren consisted of 962 participants, and thus included a similar sample size compared with the sample size of Europeans in our study. However, since this study was conducted in a European cohort, it would not have observed associations observed in individuals of African descent. Further, the majority of IFNW1 SNPs (16/17) were imputed and located outside the coding region. These SNPs would therefore not be included in the exome sequencing used by McLaren et al. and cannot be directly compared with their study. Since the SKAT-O outcome of IFNW1 was not validated in the FIRST study or elsewhere, it is important to emphasize the association of IFNW1 with HIV-1 VL reported here is lacking validation and should be interpretated cautiously. Without validation, analysis of selected candidate genes or a candidate pathway is at risk of type 1 error [60]. Therefore, an investigation of whether the association of gene level variation in IFNW1 with HIV VL is true shall preferably be in a cohort of participants similarly at an early stage of HIV-1 infection and including participants of African descent. A recent study did explore the genetic contribution to HIV-1 spVL in a cohort of PLWH from Africa (n = 2682) at the SNP level [61]. This study did not observe any SNP level associations outside the HLA region and the CHD1L gene. Although, this study did not assess gene level associations and used a more stringent p-value cut-off than used in our study. However, taken together, the lack of current validation at both the SNP and gene level mean that the associations observed in our study should be interpreted cautiously.

There were several limitations to this study that warrant discussion. A common limitation to genetic studies is the diversity and the frequencies of genotypes and small sample sizes across the ethnic subpopulations. In our study, and particularly in the demographically diverse START cohort, we are challenged by this. We attempted to control for potential demographic and environmental biases in three ways. Firstly, by including principal components in the main model. Second, by conducting subgroup analyses where participants are separated into more demographically homogenous groups. Thirdly, by performing sensitivity analyses that utilised linear mixed effects models to control for population stratification, in addition to principal components. These methods are suggested to be superior to the use of PCs alone to control for population structure in genetic association studies. In our results, we see that the p-value for the IFNW1 association with HIV-VL increase above the multiple testing threshold used in this study. This indicates that there may be biases introduced by population structures that account for some of the signal in the main model. Due to the limited statistical power considering fewer participants within subgroups, we did not investigate further participant subgroups based on country of enrolment. As a common challenge in genetic studies, the sample size of our cohort limits our ability to analyse rare variants, as well as providing sufficient statistical power to analyse diversity in ethnicity of study participants. Use of SNP array genotype data, as opposed to having e.g., genome wide sequencing data, which would provide further information on rare variants, is also a limitation. Further, our study did not assess type 1 IFN gene expression or serum type 1 IFN measures. Finally, it is important to note our study did not assess or incorporate knowledge of genetic variants in CCR5 (including CCR5D32) and HLA genes known to affect viral load. Thus, the association detected of IFNW1 SNPs and study entry VL may be due to variation in genes prior known to affect VL. A strength of the START study is in the cohort size of European participants.

To conclude, in cumulated SKAT-O analysis on imputed type 1 IFN SNPs from a cohort of ART naïve HIV-1 positive participants, we did not detect an association with HIV disease progression. However, for the cohort of HIV-1 positive individuals in early stages of their infection, we detected an association between IFNW1, and HIV-1 study entry VL. This association was most apparent in persons enrolled in Africa. The observed association was no longer significant after more stringent control for population structure and the results observed here should be interpreted cautiously without additional validation.

Availability of data and materials

Due to data protection regulations and patient confidentiality concerns, the datasets analyzed in this study are only available via reasonable request to the corresponding author and approval from the INSIGHT Scientific Steering Committee. Software used for this study was written in R and can be obtained at under the conditions outlined in GNU GPL-3.0 license.





Flexible initial retrovirus suppressive therapies


Genome wide association study


HIV viral load


Human leukocyte antigen




Linkage disequilibrium


Principal components analysis


People living with HIV


Quality control


Set point viral load


Sequence kernel association test, optimised variant


Single nucleotide polymorphism


Strategic timing of antiretroviral treatment trial


Viral load


  1. Mclaren PJ, Carrington M. The impact of host genetic variation on infection with HIV-1. Nat Immunol. 2015;16:577–83.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  2. Fraser C, Lythgoe K, Leventhal GE, Shirreff G, Hollingsworth TD, Alizon S, et al. Virulence and pathogenesis of HIV-1 infection: An evolutionary perspective. Science. 2014;343(6177):1–18.

    Article  Google Scholar 

  3. McLaren PJ, Coulonges C, Bartha I, Lenz TL, Deutsch AJ, Bashirova A, et al. Polymorphisms of large effect explain the majority of the host genetic contribution to variation of HIV-1 virus load. Proc Natl Acad Sci. 2015;112(47):14658–63.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  4. Gabrielaite M, Bennedbæk M, Zucco AG, Ekenberg C, Murray DD, Kan VL, et al. Human immunotypes impose selection on viral genotypes through viral epitope specificity. J Infect Dis. 2021;224(12):2053–63.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  5. Marmor M, Hertzmark K, Thomas SM, Halkitis PN, Vogler M. Resistance to HIV infection. J Urban Heal. 2006;83(1):5–17.

    Article  CAS  Google Scholar 

  6. Fellay J, Ge D, Shianna KV, Colombo S, Ledergerber B, Cirulli ET, et al. Common genetic variation and the control of HIV-1 in humans. PLoS Genet. 2009;5(12):e1000791.

    Article  PubMed  PubMed Central  Google Scholar 

  7. McLaren PJ, Pulit SL, Gurdasani D, Bartha I, Shea PR, Pomilla C, et al. Evaluating the impact of functional genetic variation on HIV-1 control. J Infect Dis. 2017;216(9):1063–9.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  8. Tough RH, McLaren PJ. Interaction of the host and viral genome and their influence on HIV disease. Front Genet. 2019;10:1–9.

    Google Scholar 

  9. Ohainle M, Helms L, Vermeire J, Roesch F, Humes D, Basom R, et al. A virus-packageable CRISPR screen identifies host factors mediating interferon inhibition of HIV. Elife. 2018;7:1–32.

    Article  Google Scholar 

  10. Wang K, Li M, Hakonarson H. Analysing biological pathways in genome-wide association studies. Nat Rev Genet. 2010;11(12):843–54.

    Article  CAS  PubMed  Google Scholar 

  11. Pairo-Castineira E, Clohisey S, Klaric L, Bretherick AD, Rawlik K, Pasko D, et al. Genetic mechanisms of critical illness in COVID-19. Nature. 2021;591(7848):92–8.

    Article  PubMed  Google Scholar 

  12. Al QZ, et al. Inborn errors of type I IFN immunity in patients with life-threatening COVID-19. Science. 2020;370:4570.

    Article  Google Scholar 

  13. Diop G, Hirtzig T, Do H, Coulonges C, Vasilescu A, Labib T, et al. Exhaustive genotyping of the interferon alpha receptor 1 (IFNAR1) gene and association of an IFNAR1 protein variant with AIDS progression or susceptibility to HIV-1 infection in a French AIDS cohort. Biomed Pharmacother. 2006;60(9):569–77.

    Article  CAS  PubMed  Google Scholar 

  14. Sandstrom TS, Ranganath N, Angel JB. Impairment of the type I interferon response by HIV-1: Potential targets for HIV eradication. Cytokine Growth Factor Rev. 2017;2017(37):1–16.

    Article  Google Scholar 

  15. Soper A, Kimura I, Nagaoka S, Konno Y, Yamamoto K, Koyanagi Y, et al. Type i interferon responses by HIV-1 infection: association with disease progression and control. Front Immunol. 2018;5(8):1–11.

    Article  Google Scholar 

  16. Doyle T, Goujon C, Malim MH. HIV-1 and interferons: Who’s interfering with whom? Nat Rev Microbiol. 2015;13(7):403–13.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  17. Bergantz L, Subra F, Deprez E, Delelis O, Richetta C. Interplay between intrinsic and innate immunity during HIV infection. Cells. 2019.

    Article  PubMed  PubMed Central  Google Scholar 

  18. Yin X, Langer S, Zhang Z, Herbert KM, Yoh S. Sensor sensibility—HIV-1 and the innate immune response. Cells. 2019;9:1–18.

    Google Scholar 

  19. Gingras SN, Tang D, Tuff J, McLaren PJ. Minding the gap in HIV host genetics: opportunities and challenges. Human Genet. 2020;139:865–75.

    Article  Google Scholar 

  20. Touloumi G, Pantazis N, Babiker AG, Walker SA, Katsarou O, Karafoulidou A, et al. Differences in HIV RNA levels before the initiation of antiretroviral therapy among 1864 individuals with known HIV-1 seroconversion dates. AIDS. 2004;18(12):1697–705.

    Article  CAS  PubMed  Google Scholar 

  21. Ekenberg C, Tang MH, Zucco AG, Murray DD, MacPherson CR, Hu X, et al. Association between single-nucleotide polymorphisms in hla alleles and human immunodeficiency virus type 1 viral load in demographically diverse, antiretroviral therapy-naive participants from the strategic timing of antiretroviral treatment trial. J Infect Dis. 2019;220(8):1325–34.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  22. MacArthur RD, Novak RM, Peng G, Chen L, Xiang Y, Hullsiek KH, et al. A comparison of three highly active antiretroviral treatment strategies consisting of non-nucleoside reverse transcriptase inhibitors, protease inhibitors, or both in the presence of nucleoside reverse transcriptase inhibitors as initial therapy (CPCRA 05. Lancet. 2006;368(9553):2125–35.

    Article  CAS  PubMed  Google Scholar 

  23. Initiation of Antiretroviral Therapy in Early Asymptomatic HIV Infection. 2015. N Engl J Med. 373(9): 795–807.

  24. Murray DD, Grund B, Macpherson CR, Ekenberg C, Zucco AG, Reekie J, et al. Association between ten-eleven methylcytosine dioxygenase 2 genetic variation and viral load in people with HIV. AIDS. 2023;37(3):379–87.

    Article  CAS  PubMed  Google Scholar 

  25. Jassal B, Matthews L, Viteri G, Gong C, Lorente P, Fabregat A, et al. The reactome pathway knowledgebase. Nucleic Acids Res. 2019.

    Article  PubMed Central  Google Scholar 

  26. Gillespie M, Jassal B, Stephan R, Milacic M, Rothfels K, Senff-Ribeiro A, et al. The reactome pathway knowledgebase 2022. Nucleic Acids Res. 2022;50(D1):D687-92.

    Article  CAS  PubMed  Google Scholar 

  27. Hunt SE, McLaren W, Gil L, Thormann A, Schuilenburg H, Sheppard D, et al. Ensembl variation resources. Database. 2018.

    Article  PubMed  PubMed Central  Google Scholar 

  28. Cunningham F, Allen JE, Allen J, Alvarez-Jarreta J, Amode MR, Armean IM, et al. Ensembl. Nucleic Acids Res. 2022;50(D1):988–95.

    Article  Google Scholar 

  29. Smedley D, Haider S, Ballester B, Holland R, London D, Thorisson G, et al. BioMart—biological queries made easy. BMC Genomics. 2009;10(1):22.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  30. Auton A, Abecasis GR, Altshuler DM, Durbin RM, Abecasis GR, Bentley DR, et al. A global reference for human genetic variation. Nature. 2015;526(7571):68–74.

    Article  PubMed  Google Scholar 

  31. Howie B, Fuchsberger C, Stephens M, Marchini J, Abecasis GR. Fast and accurate genotype imputation in genome-wide association studies through pre-phasing. Nat Genet. 2013;44(8):955–9.

    Article  Google Scholar 

  32. Chang CC, Chow CC, Tellier LCAM, Vattikuti S, Purcell SM, Lee JJ. Second-generation PLINK: rising to the challenge of larger and richer datasets. Gigascience. 2015;4(1):1–16.

    Article  Google Scholar 

  33. Delaneau O, Zagury JF, Marchini J. Improved whole-chromosome phasing for disease and population genetic studies. Nat Methods. 2013;10(1):5–6.

    Article  CAS  PubMed  Google Scholar 

  34. Alexander DH, Novembre J, Lange K. Fast model-based estimation of ancestry in unrelated individuals. Genome Res. 2009;19(9):1655–64.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  35. Price AL, Patterson NJ, Plenge RM, Weinblatt ME, Shadick NA, Reich D. Principal components analysis corrects for stratification in genome-wide association studies. Nat Genet. 2006;38(8):904–9.

    Article  CAS  PubMed  Google Scholar 

  36. Kao PYP, Leung KH, Chan LWC, Yip SP, Yap MKH. Pathway analysis of complex diseases for GWAS, extending to consider rare variants, multi-omics and interactions. Biochim Biophys Acta Gen Subj. 2017;1861(2):335–53.

    Article  CAS  PubMed  Google Scholar 

  37. Lee J, Kim YJ, Lee J, Kim BJ, Lee S, Park T. Gene-set association tests for next-generation sequencing data. Bioinformatics. 2016;32(17):i611–9.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  38. Lee S, Wu MC, Lin X. Optimal tests for rare variant effects in sequencing association studies. Biostatistics. 2012;13(4):762–75.

    Article  PubMed  PubMed Central  Google Scholar 

  39. Wu MC, Lee S, Cai T, Li Y, Boehnke M, Lin X. Rare-variant association testing for sequencing data with the sequence kernel association test. Am J Hum Genet. 2011;89(1):82–93.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  40. Ionita-Laza I, Lee S, Makarov V, Buxbaum JD, Lin X. Sequence kernel association tests for the combined effect of rare and common variants. Am J Hum Genet. 2013;92(6):841–53.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  41. Maglott D, Ostell J, Pruitt KD, Tatusova T. Entrez gene: gene-centered information at NCBI. Nucl Acids Res. 2007;35(SUPPL. 1):26–31.

    Article  Google Scholar 

  42. Lee S (Shawn). SKAT Package. 2022.

  43. SKAT Download.

  44. Mbatchou J, Barnard L, Backman J, Marcketta A, Kosmicki JA, Ziyatdinov A, et al. Computationally efficient whole-genome regression for quantitative and binary traits. Nat Genet. 2021;53(7):1097–103.

    Article  CAS  PubMed  Google Scholar 

  45. Schlusser KE, Sharma S, de la Torre P, Tambussi G, Draenert R, Pinto AN, et al. Comparison of self-report to biomarkers of recent hiv infection: findings from the start trial. AIDS Behav. 2018;22(7):2277–83.

    Article  PubMed  PubMed Central  Google Scholar 

  46. The major genetic determinants of HIV-1 Control Affect HLA Class I Peptide Presentation. Science. 2010. 330(6010): 1551–7.

  47. Sherry ST. dbSNP: the NCBI database of genetic variation. Nucleic Acids Res. 2001;29(1):308–11.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  48. Landrum MJ, Lee JM, Benson M, Brown GR, Chao C, Chitipiralla S, et al. ClinVar: improving access to variant interpretations and supporting evidence. Nucl Acids Res. 2018;46(D1):D1062-7.

    Article  CAS  PubMed  Google Scholar 

  49. Li S, Zhao F, Shao J, Xie Y, Chang H, Zhang Y. Interferon-omega: current status in clinical applications Shi-fang. Int Immunopharmacol. 2017;52:253–60.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  50. Künzi MS, Pitha PM. Role of interferon-stimulated gene ISG-15 in the interferon-ω-mediated inhibition of human immunodeficiency virus replication. J Interf Cytokine Res. 1996;16(11):919–27.

    Article  Google Scholar 

  51. PubMed.

  52. Refae S, Gal J, Ebran N, Otto J, Borchiellini D, Peyrade F, et al. Germinal Immunogenetics predict treatment outcome for PD-1/PD-L1 checkpoint inhibitors. Invest New Drugs. 2020;38(1):160–71.

    Article  CAS  PubMed  Google Scholar 

  53. Lenci RE, Bevier M, Brandt A, Bermejo JL, Sucker A, Moll I, et al. Influence of genetic variants in type i interferon genes on melanoma survival and therapy. PLoS One. 2012;7(11):e50692.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  54. Choi J, Song N, Han S, Chung S, Sung H, Lee JY, et al. The associations between immunity-related genes and breast cancer prognosis in Korean women. PLoS One. 2014;9(7):e103593.

    Article  PubMed  PubMed Central  Google Scholar 

  55. Lu S, Pardini B, Cheng B, Naccarati A, Huhn S, Vymetalkova V, et al. Single nucleotide polymorphisms within interferon signaling pathway genes are associated with colorectal cancer susceptibility and survival. PLoS One. 2014;9(10):e111061.

    Article  PubMed  PubMed Central  Google Scholar 

  56. gnomAD.

  57. Napravnik S, Poole C, Thomas JC, Eron JJ. Gender difference in HIV RNA levels: a meta-analysis of published studies. J Acquir Immune Defic Syndr. 2002;31(1):11–9.

    Article  CAS  PubMed  Google Scholar 

  58. Rindler AE, Kusejko K, Kuster H, Neumann K, Leemann C, Zeeb M, et al. The interplay between replication capacity of HIV-1 and surrogate markers of disease. J Infect Dis. 2022;226(6):1057–68.

    Article  CAS  PubMed  Google Scholar 

  59. Petersdorf N, Ross JM, Weiss HA, Barnabas RV, Wasserheit JN, Beguelin C, et al. Systematic review and meta-analysis of hepatitis C virus infection and HIV viral load: new insights into epidemiologic synergy. J Int AIDS Soc. 2016;19:1–8.

    Article  Google Scholar 

  60. Clohisey S, Baillie JK. Host susceptibility to severe influenza A virus infection. Crit Care. 2019;23(1):1–10.

    Article  Google Scholar 

  61. McLaren PJ, Porreca I, Iaconis G, Mok HP, Mukhopadhyay S, Karakoc E, et al. Africa-specific human genetic variation near CHD1L associates with HIV-1 load. Nature. 2023;620(7976):1025–30.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

Download references


New England Journal of Medicine 2015;373:795–807 includes the complete list of START investigators. Lancet 2006;368:2125–35 includes the complete list of FIRST investigators. We acknowledge all study participants and colleagues that have contributed to participant inclusion and collection of data as part of the clinical trials.


Open access funding provided by Copenhagen University. The Danish National Research Foundation funded this work (DNRF126). START and FIRST were supported by the National Institutes of Health (UM1-AI068641, UM1-AI120197, U01-AI36780, U01-AI042170, U01- AI046362), National Institute of Allergy and Infectious Diseases, National Institutes of Health Clinical Center, National Cancer Institute, National Heart, Lung, and Blood Institute, Eunice Kennedy Shriver National Institute of Child Health and Human Development, National Institute of Mental Health, National Institute of Neurological Disorders and Stroke, National Institute of Arthritis and Musculoskeletal and Skin Diseases, Agence Nationale de Recherches sur le SIDA et les Hepatites Virales (France), National Health and Medical Research Council (Australia), National Research Foundation (Denmark), Bundes ministerium für Bildung und Forschung (Germany), European AIDS Treatment Network, Medical Research Council (United Kingdom), National Institute for Health Research, National Health Service (United Kingdom), and University of Minnesota. Antiretroviral drugs donations to the central drug repository were provided by AbbVie, Bristol-Myers Squibb, Gilead Sciences, GlaxoSmithKline/ViiV Healthcare, Janssen Scientific Affairs, and Merck. Bristol Myer Squibb and GlaxoSmithKline contributed antiretroviral drugs to the FIRST study.

Author information

Authors and Affiliations




DDM, JR, JL: Planned the study. SBM, DDM, PL, JR, JL: Contributed to the design. CRM, PL and MM: Implemented and modified the bioinformatics algorithms. All authors: contributed to the interpretation of results. SBM, PL, JR, DDM, MH, LB, JL: Wrote the manuscript. All authors helped in the interpretation of results as well as in critical review and approval of the final version of the manuscript.

Corresponding author

Correspondence to Daniel D. Murray.

Ethics declarations

Ethics approval and consent to participate

All participants gave written consent to participate in research and the comprehensive genetic analysis. All participating sites have received approvals from their Institutional Review Board (IRB) or Ethics Committee (EC) prior to implementation. The confidentiality of study participants will be protected in accordance with standard IRB/EC policies and procedures.

Consent for publication

Not applicable.

Competing interests

No competing interests for this work were declared.

Additional information

Publisher's Note

Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Supplementary Information

Additional file 1.

Pairwise principal component plots (PCs 1-5) illustrating population stratification for (A) START and (B) FIRST cohort. The START cohort are overlayed with ancestry estimations.

Additional file 2.

Table of SNPs in type 1 IFN Genes (N=673).

Additional file 3.

Linkage Disequilibrium Heatmap values for Fig. 1.

Rights and permissions

Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit The Creative Commons Public Domain Dedication waiver ( applies to the data made available in this article, unless otherwise stated in a credit line to the data.

Reprints and permissions

About this article

Check for updates. Verify currency and authenticity via CrossMark

Cite this article

Mørup, S.B., Leung, P., Reilly, C. et al. The association between single-nucleotide polymorphisms within type 1 interferon pathway genes and human immunodeficiency virus type 1 viral load in antiretroviral-naïve participants. AIDS Res Ther 21, 27 (2024).

Download citation

  • Received:

  • Accepted:

  • Published:

  • DOI: