HIV Peptides Research Articles

Comprehensive analysis of HIV Gag‐specific IFN‐γ response in HIV‐1‐ and HIV‐2‐infected asymptomatic patients from a clinical cohort in The Gambia

Majority of HIV-2-infected individuals meet the criteria of long-term non-progressors. This has been linked to superior qualitative HIV-2-specific cellular immune responses that correlate with viral control. However, it is unknown whether this is due to frequent targeting of immunodominant Gag epitopes in HIV-2 than HIV-1 infection. We describe a comprehensive comparison of the magnitude, breadth and frequency of Gag responses and the degree of cross-recognition of frequently targeted, immunodominant Gag peptides in a cross-sectional study of asymptomatic HIV-1- and HIV-2-infected individuals. Fresh PBMC from 20 HIV-1- and 20 HIV-2-infected patients with similar CD4(+) T-cell counts (p=0.36) were stimulated with pools of HIV-1 and/or HIV-2 Gag peptides in an IFN-gamma ELISPOT assay. We found no difference in the cumulative magnitude of IFN-gamma responses (p=0.75) despite significantly lower plasma viral loads in HIV-2-infected people (p<0.0001). However, Gag211-290 was targeted with significantly higher magnitude in HIV-2-infected subjects (p=0.03) although this did not correlate with viral control. There was no difference in frequently targeted Gag peptides, the breadth, immunodominance or cross-recognition of Gag peptide pools between the two infections. This suggests that other factors may control viral replication in HIV-2 infection.

Optimizing peptide matrices for identifying T‐cell antigens

Mapping T-cell epitopes for a pathogen or vaccine requires a complex method for screening hundreds to thousands of peptides with a limited amount of donor sample. We describe an optimized deconvolution process by which peptides are pooled in a matrix format to minimize the number of tests required to identify peptide epitopes. Four peptide pool matrices were constructed to deconvolute the HIV-specific T-cell response in three HIV-infected individuals. ELISpot assays were used to map peptide antigens. Many HIV peptides were mapped in all three individuals. However, there were several challenges and limitations associated with the deconvolution process. Peptides that induced low-frequency responses or were masked by peptide competition within a given pool were not identified, because they did not meet the threshold criteria for a positive response. Also, amino acid sequence variation limited the ability of this method to map autologous HIV peptides. Alternative analysis strategies and revisions to the original matrix optimizations are presented that address ways to increase peptide identification. This optimized deconvolution method allows for efficient mapping of T-cell peptide epitopes. It is rapid, powerful, efficient, and unrestricted by HLA type.

The TRAIL of Helpless CD8+T Cells in HIV Infection

Our understanding of how CD4(+) T cells can regulate CD8(+) T cell responses in HIV infection is still incomplete. Recent evidence obtained in mice suggests that CD4(+) T cell help is required for efficient CD8(+) T cell-mediated immunity in chronic infection: CD8(+) T cells primed in the absence of such help release the TNF-related apoptosis-inducing ligand TRAIL and undergo apoptosis. Using a novel ELISPOT assay, in the present study we show that CD8(+) T cells are also a source of the antigen-specific TRAIL response in HIV-infected patients with CD4(+) T cell counts below 200. In patients with CD4(+) T cell counts above 200 TRAIL was not detectable. Accordingly, antigens to which patients have likely been exposed when CD4(+) T cell levels were high (e.g., influenza, CMV, and EBV) did not induce TRAIL. Within the HIV-positive donor population with low CD4(+) T cell counts a dissociation of the interferon-gamma (IFN-gamma) and TRAIL response to different HIV peptide epitopes was detectable suggesting impaired immunity to antigens that triggered TRAIL in the absence of IFN-gamma. Our findings emphasize that "helpless" CD8(+) T cells, i.e., cells that have been primed in the absence of CD4(+) T cell help, may play a crucial role in HIV infection. A "helpless" phenotype may impair CD8(+) T cell control of HIV and other infections and possibly contribute to the depletion of CD4(+) T cells via apoptosis. Immunizations and infections in this "helpless" state might result in ineffective CD8(+) T cell responses.

Evaluation of recombinant Onchocerca volvulus activation associated protein-1 (ASP-1) as a potent Th1-biased adjuvant with a panel of protein or peptide-based antigens and commercial inactivated vaccines

Alum, the only adjuvant approved for clinical applications, can induce strong humoral (Th2) but weak cellular (Th1) immune responses. It is necessary to develop safe and effective adjuvants capable of inducing both humoral and cellular immune responses. We previously showed that activation-associated protein-1 (ASP-1) derived from Onchocerca volvulus has potent adjuvant activity. In this study, we have further evaluated the adjuvanticity of recombinant ASP-1 using a panel of recombinant proteins or synthetic peptide-based antigens, including ovalbumin (OVA), synthetic HIV peptide (HIV-p), recombinant HIV gp41 (rgp41) and HBV HBsAg, as well as three commercially available inactivated vaccines against haemorrhagic fever with renal syndrome (HFRS), Influenza and Rabies. Our results indicate that ASP-1 induced significantly higher IgG1 (Th2-associated) and IgG2a (Th1-associated) responses than alum adjuvant against OVA antigen, HIV-p, and rgp41. Consistently, it induced similar level of IgG1 responses as alum but higher level of IgG2a and IFN-γ-producing T cell responses than alum adjuvant against HBsAg. Further, ASP-1 improved both IgG1 and IgG2a responses to three commercial inactivated vaccines when used separately or in combination. In conclusion, the recombinant ASP-1, unlike alum adjuvant, is able to induce both Th1 and Th2-associated humoral responses and Th1 cellular responses, suggesting that it can be further developed as a promising adjuvant for subunit-based and inactivated vaccines.

A gp120 HIV peptide with high similarity to HLA class II beta chains enhances PPD-specific and autoreactive T cell activation.

The recent report that anti-gp120 antibodies can be induced by allogeneic stimuli in experimental animals in the absence of HIV, has focused attention on the structural similarities between gp120 and MHC. Here we report that some HIV+ individuals develop antibodies which similarly react with the gp120 HIV sequence (aa 254-263) and with the HLA-DR beta chains (aa 142-151). As these two peptides share a high level of similarity, we have investigated the role of this gp120 region on HLA class II mediated T cell recognition. The synthetic peptide corresponding to the gp120 HIV sequence aa 254-263 has been tested on T cell line (TCL) activation. Both the PPD-specific and the self-HLA reactive TCL proliferation increased in the presence of this peptide. Prepulsing experiments indicate that this enhancing effect carried out by HIV peptide is exerted at the level of antigen presentation. Moreover, the specificity of this interaction is supported by the fact that a MoAb specific for this HIV peptide blocked the autoreactive TCL proliferation, similarly to the inhibition carried out by anticlass II antibody. These data support the hypothesis that the functional homology between the HIV peptide and the HLA beta chain described may be involved in the pathogenesis of AIDS.

Blockade of immunoregulatory Fc-signalling by HIV peptides: oligopeptides from HIV gp120 and gp41 bind the Fc portion of IgG and increase the in vitro anti-ssDNA response.

Concomitant ligation of antigen receptors with Fc-receptors negatively signals B cells. Antibodies to the Fc portion of IgG prevent this negative Fc-signalling, provided that these antibodies do not emit Fc signals. Prevention of Fc signals leads to augmented antibody responses to self and foreign antigens, and reduces the requirement for T cells by 10- to 100-fold in T cell-dependent antibody responses. In ELISA assays, peptides from conserved portions of the glycoproteins, HIV-1 gp120 or gp41 from HIV-1 and HIV-2 bind to the Fc portion of IgG, but do not bind the F(ab')2 portion of IgG. HIV-derived peptides, which bind to the Fc portion of IgG, augment the antibody-forming cell response to single-stranded (ss)DNA. The spontaneous response to ssDNA using spleen cells from young mice, and the response in the presence of exogenous DNA using spleen cells from old mice, are augmented to the greatest extent. These results demonstrate that HIV peptides bind to the Fc portion of IgG and augment immune responses to DNA; they suggest the possibility that blockade of the Fc portion of IgG antibodies is associated with a reduction in Fc-mediated regulation of anti-self responses. Blockade of regulatory Fc-signalling may account for increased circulating immunoglobulins and autoantibodies in clinical AIDS.

Higher Homologous and Lower Cross-Reactive Gag-Specific T-Cell Responses in Human Immunodeficiency Virus Type 2 (HIV-2) Than in HIV-1 Infection

Human immunodeficiency virus type 2 (HIV-2) infection results in slower CD4(+) T-cell decline, lower plasma viral load levels, and hence slower progression of the disease than does HIV-1 infection. Although the reasons for this are not clear, it is possible that HIV-2 replication is more effectively controlled by host responses. We used aligned pools of overlapping HIV-1 and HIV-2 Gag peptides in an enhanced gamma interferon enzyme-linked immunospot assay to compare the levels of homologous and cross-reactive Gag-specific T-cell responses between HIV-1- and HIV-2-infected patients. HIV-2-infected patients showed broader and stronger homologous Gag-specific T-cell responses than HIV-1-infected patients. In contrast, the cross-reactive T-cell responses in HIV-2-infected patients were both narrower and weaker than those in HIV-1-infected patients, in line with overall weaker correlations between homologous and heterologous T-cell responses among HIV-2-infected patients than among HIV-1-infected patients. Cross-reactive responses in HIV-2-infected patients tended to correlate directly with HIV-1/HIV-2 Gag sequence similarities; this was not found in HIV-1-infected patients. The CD4(+) T-cell counts of HIV-2-infected patients correlated directly with homologous responses and inversely with cross-reactive responses; this was not found in HIV-1-infected patients. Our data support a model whereby high-level HIV-2-specific T-cell responses control the replication of HIV-2, thus limiting viral diversification and priming of HIV-1 cross-reactive T-cell responses over time. However, we cannot exclude the possibility that HIV-2 replication is controlled by other host factors and that HIV-2-specific T-cell responses are better maintained in the context of slow viral divergence and a less damaged immune system. Understanding the nature of immune control of HIV-2 infection could be crucial for HIV vaccine design.

Identification of critical antibody-binding sites in the HIV-1 gp41 six-helix bundle core as potential targets for HIV-1 fusion inhibitors

Formation of the six-helix bundle (6-HB) core between the N- and C-terminal heptad repeats (NHR and CHR) regions of the HIV-1 envelope glycoprotein (Env) transmembrane subunit gp41 is a critical step during the process of virus and target cell membrane fusion. In the present study, we generated a panel of five monoclonal antibodies (mAbs) which specifically recognized the HIV-1 gp41 6-HB formed by the NHR-peptide N36 and CHR-peptide C34 mixture, but did not react with the isolated peptides N36 and C34. These mAbs did not block the HIV-1 Env-mediated cell–cell fusion at physiological temperature (37 °C), but inhibited the HIV-1 Env-mediated cell–cell fusion at suboptimal temperature (31.5 °C), under which condition the fusion process is slowed down and the viral 6-HB becomes accessible. The fusion inhibitory activity of the mAbs is correlated with their binding affinity with the 6-HB core. By screening 24 6-HB variants with single mutations at the b, c, and f positions in the helical wheels, we found that the critical binding sites of these mAbs were localized in the N-terminal region of the NHR and the C-terminal region of the CHR. These sites may serve as targets for design of small molecule HIV fusion inhibitors, e.g., organic compounds, peptides, and low molecular weight proteins.

T cell responses generated by HIV vaccines in clinical trials

Here we summarize evaluations of T cell immune responses in HIV vaccine clinical trials, highlighting investigations addressing optimization and validation of assays, reagents for determining multiclade HIV-specific responses, and current findings in phase I and II clinical testing. Maintaining peripheral blood mononuclear cell quality and function is critical for optimal detection of T cell responses. Two recent reports emphasize the necessity for processing and freezing peripheral blood mononuclear cell within 8-12 h of venipuncture at local processing sites. Analytical designs of HIV peptide panels for use in T cell assays permit standardization of measurements between vaccines and laboratories, and evaluation of responses representing global HIV-1 strains. Phase I trials indicate superior induction of T cell response frequencies and magnitude using recombinant DNA and adenovirus serotype 5 (Ad5) vaccine with HIV-1 gene inserts. Additionally, use of a unique adjuvant with a recombinant Nef/Tat/gp120 vaccine induced high titer anti-HIV neutralizing antibodies, T cell proliferation, but not CD8 T cell responses. T cell functional assays have achieved the standardization and validation to support licensure of vaccine candidates. Studies in progress will determine if these measurements are critical in defining correlates of immune protection.

Enhancement of epitope-specific cellular immune responses by immunization with HIV-1 peptides genetically conjugated to the B-subunit of recombinant cholera toxin

As more HIV-1 infected patients receive anti-retroviral drug treatment, the occurrence of drug-resistant variants of the virus is increasing. We have previously shown that mutated HIV peptide sequences represent mutations induced by antiretroviral drugs are equally good and often better immunogens than wild type peptides. The non-toxic B subunit of cholera toxin (CTB) is an active substance in the oral cholera vaccine, and has been shown to bind ganglioside receptors and activate mucosal cells. By fusing mutant epitopes deriving from HIV-1 enzymes with the B subunit of cholera toxin, we aim is to induce cellular responses against virus harboring drug-induced mutations. We successfully created conjugates of HIV peptide sequences fused to rCTB. The immune response against the different peptides was strongly enhanced by the fusion to the toxin. Moreover, immunization with sequence containing drug-induced mutation triggered a cross-reactive immune response against the wild type epitope. Long-term follow-up of immunized animals revealed a persistence of cellular immune response for over 4 months, which could readily be boosted with an additional late immunization. By linking HIV-peptides to the B subunit of cholera toxin it is thus possible to stimulate a strong and long-lasting immune response, significantly better than that evoked by the peptide alone.

Transmission of HIV-1 CTL Escape Variants Provides HLA-Mismatched Recipients with a Survival Advantage

One of the most important genetic factors known to affect the rate of disease progression in HIV-infected individuals is the genotype at the Class I Human Leukocyte Antigen (HLA) locus, which determines the HIV peptides targeted by cytotoxic T-lymphocytes (CTLs). Individuals with HLA-B*57 or B*5801 alleles, for example, target functionally important parts of the Gag protein. Mutants that escape these CTL responses may have lower fitness than the wild-type and can be associated with slower disease progression. Transmission of the escape variant to individuals without these HLA alleles is associated with rapid reversion to wild-type. However, the question of whether infection with an escape mutant offers an advantage to newly infected hosts has not been addressed. Here we investigate the relationship between the genotypes of transmitted viruses and prognostic markers of disease progression and show that infection with HLA-B*57/B*5801 escape mutants is associated with lower viral load and higher CD4+ counts.

Membrane-specific antibodies induced by liposomes can simultaneously bind to HIV-1 protein, peptide, and membrane lipid epitopes

Liposomes containing lipid A as an adjuvant and also containing either both (a) cholesterol and gp140 HIV envelope protein or (b) galactosylceramide and a 48 amino acid peptide from the membrane proximal external region of gp41 from HIV, as liposomal antigens were used for immunizing mice. Monoclonal antibodies from each type of immunization were obtained, which recognized either the lipid antigen or the protein (or peptide) antigen separately, or that simultaneously bound to both the lipid and protein (or peptide) antigens, by ELISA. After immunizing with liposomes containing both phosphatidylinositol 4-phosphate (PIP) and peptide antigen, a unique monoclonal antibody was also obtained, which did not bind separately either to the lipid or peptide antigen, or to the liposomes alone, but did bind to the original liposomal antigen containing the peptide but lacking PIP. Our data suggest that immunization with liposomes containing lipid A and also containing both a lipid and protein (or peptide) antigen induces antibodies that recognize broad topographical antigenic liposomal membrane surface patterns. These membrane-specific antibodies have unique binding characteristics.

Immunogenically optimized peptides derived from natural mutants of HIV CTL epitopes and peptide combinatorial libraries

Two strategies were aimed at identifying immunogenically optimized peptides for the potential use in the formulation of an effective prophylactic or therapeutic HIV-1 vaccine. Three CTL epitopes were investigated: Gag p24(19-27) TV9, Gag p17(77-85) SL9, and RT(309-317) IV9. The first strategy derives from the hypothesis that a number of rare mutant CTL epitopes of HIV-1 may be more immunogenic than the common ones. As such, these rare mutant sequences might be highly effective in generating cross reactive anti-HIV-1 CTL responses against a range of mutant sequences. As anticipated, several rare mutant peptide sequences were identified that generated strong CTL responses against both the consensus sequences and several naturally occurring mutants in human PBL cultures primed ex vivo and in HLA-A2 transgenic mice immunized in vivo. Finally, to reach beyond the sequence diversity of the "natural" library of mutated sequences, a synthetic combinatorial peptide library was screened with a TV9 specific T-cell line; this resulted in the identification of an immunogenically optimized mimic peptide sequence that provoked highly effective CTL immune responses against TV9 and mutants. Sequence homologies between the natural mutants and synthetic mimic may provide insight into key contact positions in the MHC/TCR/peptide complex.

Different immunodominance of HIV-1-specific CTL epitopes among three subtypes of HLA-A∗26 associated with slow progression to AIDS

It is speculated that HLA-A∗26-restricted HIV-1-specific CTLs can control HIV-1, since HLA-A∗26 is associated with a slow progression to AIDS. In three major HLA-A∗26 subtypes, HLA-A∗2601-restricted, and HLA-A∗2603-restricted HIV-1 epitopes have been identified, but HLA-A∗2602-restricted ones have not. We here identified HLA-A∗2602-restricted HIV-1 epitopes by using reverse immunogenetics and compared the immunodominance of the epitopes among the three subtypes. Out of 110 HIV-1 peptides carrying HLA-A∗26 anchor residues, only the Gag169-177 peptide, which had been previously identified as an HLA-A∗2601- and HLA-A∗2603-restricted immunodominant epitope, induced Gag169-177-specific CD8+ T cells from only two of six HLA-A∗2602+ HIV-1-infected individuals. No difference in affinity of this epitope peptide was found among these three HLA-A∗26 subtypes, indicating that Gag169-177 was effectively presented by HLA-A∗2602 but recognized as a subdominant epitope in HIV-1-infected HLA-A∗2602+ individuals. These findings indicate different immunodominance of Gag169-177 epitope among 3 HLA-A∗26 subtypes.

The saga of MHC-bound peptides: a renaissance for antigen presentation?

In this issue of the JCI, two separate studies on MHC-bound peptides reopen the debate on the utility of peptides for the purposes of vaccination and treatment of autoimmune diseases. In the first study, by Wahlström et al., peptides bound to HLA-DR17 from bronchoalveolar lavage cells of sarcoidosis patients were analyzed in order to identify target antigens of the autoimmune response (see the related article beginning on page 3576). In the second study, by Le Gall et al., the modulation of epitope immunodominance and the processing and presentation of HIV peptides for MHC class I recognition were shown to be dependent on flanking residues that were N terminal to the natural epitopes (see the related article beginning on page 3563). Both studies highlight the tremendous therapeutic potential of MHC-bound peptides. They also emphasize that technical issues are still plaguing this field and hindering our understanding of MHC presentation in vivo.

Liposome destabilization induced by synthetic lipopeptides corresponding to envelope and non-structural domains of GBV-C/HGV virus. Conformational requirements for leakage

Liposomes have been used primarily as a model system for studying biological membranes. Numerous chemical, biochemical and biophysical methods have been used to elucidate the various aspects of the interaction between proteins or peptides and phospholipids. Having in mind the potential use of synthetic lipopeptides as antiviral therapies and aiming for a better understanding of the molecular interaction of the GBV-C/HGV with liposomes as model membranes, epitopes of GBV-C/HGV located at the E2 (99-118) and NS3(440-460) regions were selected. Peptides were modified at the N-terminus with acyl chains of different length (C 14 and C 16) yielding the corresponding myristoil and palmytoil lipopeptides. The main aim of the present study was to get insight into the membrane-interacting properties of the above-described synthetic lipopeptides and to study their inhibition of the capacity of perturbing model membranes of fusion peptide of HIV-1 using fluorescence spectroscopy. In an attempt to establish a relationship between peptide membrane activity and structure, we use Circular Dichroism (CD) and Fourier-Transform Infrared Spectroscopy (FTIR).

Non-coding transcript in T cells (NTT): Antisense transcript activates PKR and NF-κB in human lymphocytes

T cell activation is a complex process involving many steps and the role played by the non-protein-coding RNAs (ncRNAs) in this phenomenon is still unclear. The non-coding T cells transcript (NTT) is differentially expressed during human T cells activation, but its function is unknown. Here, we detected a 426 nt NTT transcript by RT–PCR using RNA of human lymphocytes activated with a synthetic peptide of HIV-1. After cloning, the sense and antisense 426 nt NTT transcripts were obtained by in vitro transcription and were sequenced. We found that both transcripts are highly structured and are able to activate PKR. A striking observation was that the antisense 426 nt NTT transcript is significantly more effective in activating PKR than the corresponding sense transcript. The transcription factor NF-κB is activated by PKR through phosphorylation and subsequent degradation of its inhibitor I-κBβ. We also found that the antisense 426 nt NTT transcript induces more efficiently the degradation of I-κBβ than the sense transcript. Thus, this study suggests that the role played by NTT in the activation of lymphocytes can be mediated by PKR through NF-κB activation. However, the physiological significance of the activity of the antisense 426 nt NTT transcript remains unknown.

Cellular immune responses induced with dose-sparing intradermal administration of HIV vaccine to HIV-uninfected volunteers in the ANRS VAC16 trial.

ObjectiveThe objective was to compare the safety and cellular immunogenicity of intradermal versus intramuscular immunization with an HIV-lipopeptide candidate vaccine (LIPO-4) in healthy volunteers.MethodologyA randomized, open-label trial with 24 weeks of follow-up was conducted in France at six HIV-vaccine trial sites. Sixty-eight healthy 21– to 55–year-old HIV-uninfected subjects were randomized to receive the LIPO-4 vaccine (four HIV lipopeptides linked to a T-helper–stimulating epitope of tetanus-toxin protein) at weeks 0, 4 and 12, either intradermally (0.1 ml, 100 µg of each peptide) or intramuscularly (0.5 ml, 500 µg of each peptide). Comparative safety of both routes was evaluated. CD8+ T-cell immune responses to HIV epitopes (ELISpot interferon-γ assay) and tetanus toxin-specific CD4+ T-cell responses (lymphoproliferation) were assessed at baseline, two weeks after each injection, and at week 24.Results and ConclusionNo severe, serious or life-threatening adverse events were observed. Local pain was significantly more frequent after intramuscular injection, but local inflammatory reactions were more frequent after intradermal immunization. At weeks 2, 6, 14 and 24, the respective cumulative percentages of induced CD8+ T-cell responses to at least one HIV peptide were 9, 33, 39 and 52 (intradermal group) or 14, 20, 26 and 37 (intramuscular group), and induced tetanus toxin-specific CD4+ T-cell responses were 6, 27, 33 and 39 (intradermal), or 9, 46, 54 and 63 (intramuscular). In conclusion, intradermal LIPO-4 immunization was well tolerated, required one-fifth of the intramuscular dose, and induced similar HIV-specific CD8+ T-cell responses. Moreover, the immunization route influenced which antigen-specific T-cells (CD4+ or CD8+) were induced.Trial RegistrationClinicalTrials.gov NCT00121121

Levels of antigen processing machinery components in dendritic cells generated for vaccination of HIV-1+ subjects

To evaluate expression of the antigen processing machinery (APM) components and HLA molecules by monocyte-derived dendritic cells (DC) generated from chronically HIV-1 infected subjects on antiretroviral therapy (ART) and to assess their ability to ex vivo induce HIV-1 specific T cells. DC generated in 16 HLA-A2 positive patients were matured in cytokines, pulsed with HIV-1 or other viral peptides and tested in interferon (IFN)-gamma ELISPOT assays. Immature (i)DC, mature (m)DC and viral peptide-pulsed DC were studied by multiparameter quantitative flow cytometry for intracellular APM component expression and for HLA class I and II, beta-2 microglobulin and co-stimulatory molecule surface expression. DC from 13 normal donors served as controls. Marked heterogeneity in APM component expression levels in iDC and mDC from HIV-1 positive subjects was observed. Nevertheless, the median levels were comparable to those in iDC and mDC, respectively, from normal donors. Patients' mDC pulsed with the HIV-1, influenza A, cytomegalovirus (CMV) or Epstein-Barr virus peptides induced IFN-gamma production by T cells specific for these peptides in ELISPOT assays. The frequency of T cells responsive to influenza A, cytomegalovirus or Epstein-Barr virus peptides was comparable in the patients and normal donors. The APM component expression profiles of iDC and mDC were more heterogeneous in subjects with chronic HIV-1 infection on ART, than those in normal donors, although not statistically different. Ex vivo, patients' DC pulsed with HIV-1 peptides induced IFN-gamma production from autologous T cells. Thus, DC obtained from HIV-1 infected subjects on ART were phenotypically and functionally competent.

Leveraging hierarchical population structure in discrete association studies.

Population structure can confound the identification of correlations in biological data. Such confounding has been recognized in multiple biological disciplines, resulting in a disparate collection of proposed solutions. We examine several methods that correct for confounding on discrete data with hierarchical population structure and identify two distinct confounding processes, which we call coevolution and conditional influence. We describe these processes in terms of generative models and show that these generative models can be used to correct for the confounding effects. Finally, we apply the models to three applications: identification of escape mutations in HIV-1 in response to specific HLA-mediated immune pressure, prediction of coevolving residues in an HIV-1 peptide, and a search for genotypes that are associated with bacterial resistance traits in Arabidopsis thaliana. We show that coevolution is a better description of confounding in some applications and conditional influence is better in others. That is, we show that no single method is best for addressing all forms of confounding. Analysis tools based on these models are available on the internet as both web based applications and downloadable source code at http://atom.research.microsoft.com/bio/phylod.aspx.