FRINK Linked Data Fragments server

Matches in SemOpenAlex for { <https://semopenalex.org/work/W4237687615> ?p ?o ?g. }

• W4237687615 endingPage "1283.e15" @default.
• W4237687615 startingPage "1272" @default.
• W4237687615 abstract "•Development of a residue-centric patient MHC-I presentation score validated by MS•MHC-I genotype is associated with the appearance of specific oncogenic mutations•Oncogenic mutation frequency negatively correlates with population MHC-I presentation•Recurrent oncogenic mutations are biased toward peptides that are poorly presented MHC-I molecules expose the intracellular protein content on the cell surface, allowing T cells to detect foreign or mutated peptides. The combination of six MHC-I alleles each individual carries defines the sub-peptidome that can be effectively presented. We applied this concept to human cancer, hypothesizing that oncogenic mutations could arise in gaps in personal MHC-I presentation. To validate this hypothesis, we developed and applied a residue-centric patient presentation score to 9,176 cancer patients across 1,018 recurrent oncogenic mutations. We found that patient MHC-I genotype-based scores could predict which mutations were more likely to emerge in their tumor. Accordingly, poor presentation of a mutation across patients was correlated with higher frequency among tumors. These results support that MHC-I genotype-restricted immunoediting during tumor formation shapes the landscape of oncogenic mutations observed in clinically diagnosed tumors and paves the way for predicting personal cancer susceptibilities from knowledge of MHC-I genotype. MHC-I molecules expose the intracellular protein content on the cell surface, allowing T cells to detect foreign or mutated peptides. The combination of six MHC-I alleles each individual carries defines the sub-peptidome that can be effectively presented. We applied this concept to human cancer, hypothesizing that oncogenic mutations could arise in gaps in personal MHC-I presentation. To validate this hypothesis, we developed and applied a residue-centric patient presentation score to 9,176 cancer patients across 1,018 recurrent oncogenic mutations. We found that patient MHC-I genotype-based scores could predict which mutations were more likely to emerge in their tumor. Accordingly, poor presentation of a mutation across patients was correlated with higher frequency among tumors. These results support that MHC-I genotype-restricted immunoediting during tumor formation shapes the landscape of oncogenic mutations observed in clinically diagnosed tumors and paves the way for predicting personal cancer susceptibilities from knowledge of MHC-I genotype. Avoiding immune destruction is a hallmark of cancer (Hanahan and Weinberg, 2011Hanahan D. Weinberg R.A. Hallmarks of cancer: The next generation.Cell. 2011; 144: 646-674Abstract Full Text Full Text PDF PubMed Scopus (42748) Google Scholar), suggesting that the ability of the immune system to detect and eliminate neoplastic cells is a major deterrent to tumor progression. Indeed, recent studies have demonstrated that the immune system is capable of eliminating tumors when the mechanisms that tumor cells employ to evade detection are countered (Brahmer et al., 2012Brahmer J.R. Tykodi S.S. Chow L.Q.M. Hwu W.J. Topalian S.L. Hwu P. Drake C.G. Camacho L.H. Kauh J. Odunsi K. et al.Safety and activity of anti-PD-L1 antibody in patients with advanced cancer.N. Engl. J. Med. 2012; 366: 2455-2465Crossref PubMed Scopus (5831) Google Scholar, Hodi et al., 2010Hodi F.S. O’Day S.J. McDermott D.F. Weber R.W. Sosman J.A. Haanen J.B. Gonzalez R. Robert C. Schadendorf D. Hassel J.C. et al.Improved survival with ipilimumab in patients with metastatic melanoma.N. Engl. J. Med. 2010; 363: 711-723Crossref PubMed Scopus (11191) Google Scholar, Topalian et al., 2012Topalian S.L. Hodi F.S. Brahmer J.R. Gettinger S.N. Smith D.C. McDermott D.F. Powderly J.D. Carvajal R.D. Sosman J.A. Atkins M.B. et al.Safety, activity, and immune correlates of anti-PD-1 antibody in cancer.N. Engl. J. Med. 2012; 366: 2443-2454Crossref PubMed Scopus (9275) Google Scholar). This discovery has motivated new efforts to identify the characteristics of tumors that render them susceptible to immunotherapy (Rizvi et al., 2015Rizvi N.A. Hellmann M.D. Snyder A. Kvistborg P. Makarov V. Havel J.J. Lee W. Yuan J. Wong P. Ho T.S. et al.Cancer immunology. Mutational landscape determines sensitivity to PD-1 blockade in non-small cell lung cancer.Science. 2015; 348: 124-128Crossref PubMed Scopus (5531) Google Scholar, Rooney et al., 2015Rooney M.S. Shukla S.A. Wu C.J. Getz G. Hacohen N. Molecular and genetic properties of tumors associated with local immune cytolytic activity.Cell. 2015; 160: 48-61Abstract Full Text Full Text PDF PubMed Scopus (2098) Google Scholar). Less attention has been directed toward the role of the immune system in shaping the tumor genome prior to immune evasion; however, such early interactions may have important implications for the characteristics of the developing tumor. The theory of cancer immunosurveillance dictates that the immune system should exert a negative selective pressure on tumor cell populations through elimination of tumor cells that harbor antigenic mutations or aberrations. Under this model, tumor precursor cells with antigenic variants would be at higher risk for immune elimination and, conversely, tumor cell populations that continue to expand should be biased toward cells that avoid producing neoantigens. In model organisms, there is accumulating experimental evidence supporting that immunosurveillance sculpts the genomes of tumors through detection and elimination of cancer cells early in tumor progression (DuPage et al., 2012DuPage M. Mazumdar C. Schmidt L.M. Cheung A.F. Jacks T. Expression of tumour-specific antigens underlies cancer immunoediting.Nature. 2012; 482: 405-409Crossref PubMed Scopus (390) Google Scholar, Kaplan et al., 1998Kaplan D.H. Shankaran V. Dighe A.S. Stockert E. Aguet M. Old L.J. Schreiber R.D. Demonstration of an interferon gamma-dependent tumor surveillance system in immunocompetent mice.Proc. Natl. Acad. Sci. USA. 1998; 95: 7556-7561Crossref PubMed Scopus (1164) Google Scholar, Koebel et al., 2007Koebel C.M. Vermi W. Swann J.B. Zerafa N. Rodig S.J. Old L.J. Smyth M.J. Schreiber R.D. Adaptive immunity maintains occult cancer in an equilibrium state.Nature. 2007; 450: 903-907Crossref PubMed Scopus (1023) Google Scholar, Matsushita et al., 2012Matsushita H. Vesely M.D. Koboldt D.C. Rickert C.G. Uppaluri R. Magrini V.J. Arthur C.D. White J.M. Chen Y.-S. Shea L.K. et al.Cancer exome analysis reveals a T-cell-dependent mechanism of cancer immunoediting.Nature. 2012; 482: 400-404Crossref PubMed Scopus (896) Google Scholar, Shankaran et al., 2001Shankaran V. Ikeda H. Bruce A.T. White J.M. Swanson P.E. Old L.J. Schreiber R.D. IFNgamma and lymphocytes prevent primary tumour development and shape tumour immunogenicity.Nature. 2001; 410: 1107-1111Crossref PubMed Scopus (2089) Google Scholar). In humans, the observed frequency of neoantigens has been reported to be unexpectedly low in some tumor types (Rooney et al., 2015Rooney M.S. Shukla S.A. Wu C.J. Getz G. Hacohen N. Molecular and genetic properties of tumors associated with local immune cytolytic activity.Cell. 2015; 160: 48-61Abstract Full Text Full Text PDF PubMed Scopus (2098) Google Scholar), suggesting that immunoediting could be taking place. However, this phenomenon has been challenging to study systematically because of the temporality of tumor-immune interactions as well as the difficulty accounting for individual differences in antigen presentation. The binding affinity of the MHC-I complex for peptides is a major determinant of antigenicity and depends predominantly on three genes encoded at the human leukocyte antigen (HLA) locus on chromosome 6: HLA-A, HLA-B, and HLA-C (Sidney et al., 2008Sidney J. Peters B. Frahm N. Brander C. Sette A. HLA class I supertypes: A revised and updated classification.BMC Immunol. 2008; 9: 1Crossref PubMed Scopus (474) Google Scholar). The HLA locus is highly polymorphic, with over 10,000 distinct alleles for the three genes documented to date (Robinson et al., 2015Robinson J. Halliwell J.A. Hayhurst J.D. Flicek P. Parham P. Marsh S.G.E. The IPD and IMGT/HLA database: Allele variant databases.Nucleic Acids Res. 2015; 43: D423-D431Crossref PubMed Scopus (1447) Google Scholar). This diversity raises the possibility that the set of oncogenic mutations that create neoantigens may differ substantially among individuals. Indeed, neoantigens found to drive tumor regression in response to immunotherapy were almost always unique to the responding tumor (Lu and Robbins, 2016Lu Y.-C. Robbins P.F. Targeting neoantigens for cancer immunotherapy.Int. Immunol. 2016; 28: 365-370Crossref PubMed Scopus (38) Google Scholar). Several studies have also reported that nonsynonymous mutation burden, rather than the presence of any particular mutation, is the common factor among responsive tumors (Rizvi et al., 2015Rizvi N.A. Hellmann M.D. Snyder A. Kvistborg P. Makarov V. Havel J.J. Lee W. Yuan J. Wong P. Ho T.S. et al.Cancer immunology. Mutational landscape determines sensitivity to PD-1 blockade in non-small cell lung cancer.Science. 2015; 348: 124-128Crossref PubMed Scopus (5531) Google Scholar). The paucity of recurrent oncogenic mutations driving effective responses to immunotherapy is suggestive that these mutations may less frequently be antigenic, possibly as a result of selective pressure by the immune system during tumor development. These observations led us to hypothesize that antigenic oncogenic mutations are eliminated during the early stages of tumor development in a manner that is dependent on the subset of the oncogenic peptidome that can be presented by an individual’s MHC-I. To find evidence in support of our hypothesis, we set out to systematically characterize the interactions between patient MHC-I allele combinations and recurrent cancer mutations for thousands of tumors from The Cancer Genome Atlas (TCGA). Currently, existing state-of-the-art in silico tools allow prediction of HLA-specific MHC-I peptide binding affinities. We thus needed to first devise a score capable of estimating the qualitative likelihood of MHC-I-based presentation of sequences containing specific mutations based on peptide binding affinities while accounting for each individual’s 6 MHC alleles. We then used this score to study interactions between patient HLA alleles and the corresponding MHC-I binding affinities for over a thousand recurrent mutations in known oncogenes and tumor suppressors, which are likely to be enriched for driver mutations and other early events in cancer development (Bozic et al., 2010Bozic I. Antal T. Ohtsuki H. Carter H. Kim D. Chen S. Karchin R. Kinzler K.W. Vogelstein B. Nowak M.A. Accumulation of driver and passenger mutations during tumor progression.Proc. Natl. Acad. Sci. USA. 2010; 107: 18545-18550Crossref PubMed Scopus (548) Google Scholar, McGranahan et al., 2015McGranahan N. Favero F. de Bruin E.C. Birkbak N.J. Szallasi Z. Swanton C. Clonal status of actionable driver events and the timing of mutational processes in cancer evolution.Sci. Transl. Med. 2015; 7: 283ra54Crossref PubMed Scopus (448) Google Scholar). This analysis revealed that patient MHC-I genotypes directly influence the probability that their tumor will acquire a recurrent oncogenic mutation, providing new evidence that immunoediting of oncogenic mutations occurs in humans and setting the stage for HLA-based precision strategies in cancer prevention and immunotherapy. To study the influence of MHC-I genotype in shaping the genomes of tumors, we developed a qualitative residue-centric presentation score and evaluated its potential to predict whether a sequence containing a residue will be presented on the cell surface. The score relies on aggregating MHC-I binding affinities across possible peptides that include the residue of interest. MHC-I peptide binding affinity predictions were obtained using the NetMHCPan3.0 tool (Vita et al., 2015Vita R. Overton J.A. Greenbaum J.A. Ponomarenko J. Clark J.D. Cantrell J.R. Wheeler D.K. Gabbard J.L. Hix D. Sette A. Peters B. The Immune Epitope Database (IEDB) 3.0.Nucleic Acids Res. 2015; 43: D405-D412Crossref PubMed Scopus (711) Google Scholar), and following published recommendations (Nielsen and Andreatta, 2016Nielsen M. Andreatta M. NetMHCpan-3.0; improved prediction of binding to MHC class I molecules integrating information from multiple receptor and peptide length datasets.Genome Med. 2016; 8: 33Crossref PubMed Scopus (314) Google Scholar), peptides receiving a rank threshold <2 and <0.5 were designated MHC-I binders and strong binders respectively. For evaluation of missense mutations, we based our score on the affinities of all 38 possible peptides of length 8–11 that incorporate the amino acid position of interest (Figure 1A), while for insertions and deletions, any resulting novel peptides of length 8–11 were considered (Figure S1A).Figure S1Scoring Residue Presentation Based on Predicted Binding Affinity, Related to Figure 1Show full caption(A) The number of 8-11-mer peptides that differed from the native sequence for recurrent in-frame indels pan-cancer.(B–E) The distribution of residue-centric presentation scores for MS-observed peptides and randomly selected residues for (B) best rank, (C) summation (rank < 2), (D) summation (rank <0.5), and (E) best rank with cleavage.(F–I) The log of the ratio between the fraction of MS-observed residues and the fraction of random residues detected over regular score intervals for (F) best rank, (G) summation (rank < 2), (H) summation (rank <0.5), and (I) best rank with cleavage.(J) A ROC curve revealing the accuracy of classification for several different presentation scoring schemes.(K) A heatmap showing the AUCs for the 16 alleles for each presentation scoring scheme.View Large Image Figure ViewerDownload Hi-res image Download (PPT) (A) The number of 8-11-mer peptides that differed from the native sequence for recurrent in-frame indels pan-cancer. (B–E) The distribution of residue-centric presentation scores for MS-observed peptides and randomly selected residues for (B) best rank, (C) summation (rank < 2), (D) summation (rank <0.5), and (E) best rank with cleavage. (F–I) The log of the ratio between the fraction of MS-observed residues and the fraction of random residues detected over regular score intervals for (F) best rank, (G) summation (rank < 2), (H) summation (rank <0.5), and (I) best rank with cleavage. (J) A ROC curve revealing the accuracy of classification for several different presentation scoring schemes. (K) A heatmap showing the AUCs for the 16 alleles for each presentation scoring scheme. We evaluated several strategies for combining peptide affinities to approximate presentation of a specific residue on the cell surface using an existing dataset of peptides bound to MHC-I molecules encoded by 16 different HLA alleles in monoallelic lymphoblastoid cell lines determined using mass spectrometry (MS) (Abelin et al., 2017Abelin J.G. Keskin D.B. Sarkizova S. Hartigan C.R. Zhang W. Sidney J. Stevens J. Lane W. Zhang G.L. Eisenhaure T.M. et al.Mass spectrometry profiling of HLA-associated peptidomes in mono-allelic cells enables more accurate epitope prediction.Immunity. 2017; 46: 315-326Abstract Full Text Full Text PDF PubMed Scopus (341) Google Scholar), the most comprehensive database of cell surface presented peptides currently available. These strategies included assigning the best rank among peptides, the total number of peptides with rank <2, the total number of peptides with rank <0.5, and the best rank weighted by predicted proteasomal cleavage (Figures S1B–S1K). We then compared the ability of these scores to discriminate these MS-derived residues from a size-matched set of randomly selected residues (STAR Methods). The best rank score (Figure 1B) provided the most reliable prediction that a particular residue position would be included in a sequence presented by the MHC-I on the cell surface (Figure 1C); thus, this score was used for all subsequent analysis. Finally, to test the best rank score’s ability to assess the presentation of cancer-related mutations, we scored the set of expressed mutations in 5 cancer cell lines to predict which would be presented by an HLA-A∗02:01-derived MHC-I (Tables S1A–S1E). Unless a mutation affects an anchor position, a peptide harboring a single amino acid change has a modest impact on peptide binding affinity and should be presented on the cell surface provided that the corresponding native sequence is presented (Tables S1F–S1J). Indeed, analyzing a database of native peptides found in complex with an HLA-A∗02:01 MHC-I in these 5 cell lines, we found that across cell lines, 9.8% of mutations predicted to strongly bind and 4.0% of mutations predicted to bind an HLA-A∗02:01 MHC-I at any strength were also supported by MS-derived peptides (Figure 1D). These experimental results validate the ability of a score derived from MHC-I binding affinities to identify mutations with a higher likelihood of generating neoantigens and support the application of this score to evaluate MHC-I genotype as a determinant of the antigenic potential of recurrent mutations in tumors. To determine whether individual variation in MHC-I genotypes results in patient-level differences in the presentation of mutations in a large human cancer cohort, we called HLA alleles for patients in the TCGA. We successfully assigned HLA-A, -B, and -C allele pairs to 9,176 of 9,839 cancer patients using three algorithms (Figure S2A) (Jia et al., 2013Jia X. Han B. Onengut-Gumuscu S. Chen W.-M. Concannon P.J. Rich S.S. Raychaudhuri S. de Bakker P.I.W. Imputing amino acid polymorphisms in human leukocyte antigens.PLoS ONE. 2013; 8: e64683Crossref PubMed Scopus (429) Google Scholar, Shukla et al., 2015Shukla S.A. Rooney M.S. Rajasagi M. Tiao G. Dixon P.M. Lawrence M.S. Stevens J. Lane W.J. Dellagatta J.L. Steelman S. et al.Comprehensive analysis of cancer-associated somatic mutations in class I HLA genes.Nat. Biotechnol. 2015; 33: 1152-1158Crossref PubMed Scopus (384) Google Scholar, Szolek et al., 2014Szolek A. Schubert B. Mohr C. Sturm M. Feldhahn M. Kohlbacher O. OptiType: Precision HLA typing from next-generation sequencing data.Bioinformatics. 2014; 30: 3310-3316Crossref PubMed Scopus (352) Google Scholar). Most alleles were called from the consensus of Optitype and Polysolver, allowing only 1 disagreement out of 6 alleles (Figure S2B), and a minority of patients without exome sequencing data were called based on genotype data using SNP2HLA (STAR Methods; Figure S2C). The remaining patients were successfully called by either Optitype or Polysolver (Figure S2D; Table S2). Only 245 of the known HLA alleles were observed in TCGA patients, and few alleles were present in more than 10% patients (Figures S2E–S2G). Allele frequencies were highly correlated with reported frequencies among healthy individuals of matched ancestry but weakly to allele frequencies of other populations (Figures S2H–S2J). We did not observe any unexpected bias in allele frequencies among cancer patients relative to matched healthy populations. To represent a patient’s ability to present a particular mutation, we devised a score for aggregating the Best Rank scores across the patient’s six MHC-I alleles (Figure 2A). We selected the harmonic mean to combine the six best rank scores across the 6 alleles because it has the desirable property that it is dominated by the minimal value. Thus, the Patient Harmonic-mean Best Rank (PHBR) score is highly influenced by the best allele but also integrates information about presentation by the other alleles. To determine the performance of the PHBR score for predicting actual presentation, an independent MS data was used including 5 cell lines expressing 6 HLA alleles typed to the fourth digit (Bassani-Sternberg et al., 2015Bassani-Sternberg M. Pletscher-Frankild S. Jensen L.J. Mann M. Mass spectrometry of human leukocyte antigen class I peptidomes reveals strong effects of protein abundance and turnover on antigen presentation.Mol. Cell. Proteomics. 2015; 14: 658-673Crossref PubMed Scopus (280) Google Scholar; Figure 2B). Receiver operating characteristic (ROC) curves were constructed from the PHBR scores for each cell line as well as the aggregated scores across the cell lines (Figure 2C) and demonstrated that the PHBR score was indeed predictive of peptide presentation in a multi-allelic setting. We identified a set of 1,018 likely driver mutations based on the criteria that these mutations occur in known oncogenes and tumor suppressors (Davoli et al., 2013Davoli T. Xu A.W. Mengwasser K.E. Sack L.M. Yoon J.C. Park P.J. Elledge S.J. Cumulative haploinsufficiency and triplosensitivity drive aneuploidy patterns and shape the cancer genome.Cell. 2013; 155: 948-962Abstract Full Text Full Text PDF PubMed Scopus (476) Google Scholar) and are observed in at least 3 tumors in TCGA (STAR Methods; Tables S3A–S3E). We then constructed a matrix of PHBR scores for patients (Figure 3; rows) versus the 1,018 recurrent oncogenic mutations (Figure 3, columns). This matrix provides a high-level view of individual differences in presentation of functional mutations causally implicated in tumorigenesis. Inspection of the PHBR score matrix highlights that some recurrent oncogenic mutations are universally poorly presented by the MHC-I, while others appear to have a high likelihood of being presented in general (Tables S3F and S3G). Over 95 mutations were predicted to have PHBR scores <4 for all patients, but no mutations had PHBR scores <1 across the entire population (Figures S3A and S3B). There was no obvious clustering of mutations according to functional consequence (missense versus indel or loss of function versus gain of function; Figure 3; columns), although distinct amino acid substitutions affecting the same residue tended to be grouped. We also compared patients based on the fraction of the 1,018 recurrent oncogenic mutations that their MHC-I genotype could potentially present. Patients’ mutation coverage ranged from as high as >86% of mutations with PHBR <4 (>39% at PHBR <1) to <54% of mutations receiving a PHBR <4 (<18% at PHBR <1) (Figures S3C and S3D). If MHC-I genotype restricts the oncogenic peptidome exposed to immune surveillance, exposed mutations should be less frequently observed than masked mutations in individual tumors (Figure 4A). As an initial approach, we mapped patient mutation status onto the PHBR score matrix (Figure 3) and divided PHBR scores into two groups: those that corresponded to observed mutations and those that corresponded to unobserved mutations. Comparing PHBR scores between these groups uncovered a bias for observed mutations to have higher PHBR scores (Figure 4B), with the largest differences apparent for PHBR scores <0.5 (Figure 4C). We took two approaches to quantifying the effect of MHC-I genotype on the probability of acquiring mutations, comparing the relationship between PHBR and mutation probability within patients (rows of Figure 3) and across patients (columns of Figure 3). As the logscale PHBR score was approximately linearly related to the logit probability of a mutation (Figures S4A and S4B), we modeled their relationship with an additive logistic regression model with non-linear effects to control for variation in mutation rates among mutations and patients.Figure S4Evaluating the Association between PBR Score and Probability of Mutation, Related to Figure 4Show full caption(A and B) Non-parametric estimate of the logit-mutation probability as a function of log-PHBR scores considering mutations ≥ 5 (A) Scatterplot of logit-mutation probability versus log-PHBR. (B) GAM-estimated logit-mutation probability versus log-PHBR score.(C–F) ORs (black squares) and their 95% CIs (discontinuous lines) for acquiring a mutation displayed for all cancer types for (C) the within-residue model for mutations occurring ≥ 5 times in TCGA and for (D) the within-patient model for mutations occurring ≥ 5 times in TCGA (E) within-residue model for mutations occurring ≥ 20 times in TCGA and (F) within-patient model for mutations occurring ≥ 20 times in TCGA.(G) A ROC curve showing the accuracy of the PHBR and the PBR for classifying the extracellular presentation of a residue by a patient’s six MHC alleles. The aggregated PHBR/PBR presentation scores for 5 cell lines expressing 6 MHC alleles was compared to the PHBR/PBR scores for a random set of residues based on the same MHC alleles.(D) Error bars denote the 1.5 IQR range.View Large Image Figure ViewerDownload Hi-res image Download (PPT) (A and B) Non-parametric estimate of the logit-mutation probability as a function of log-PHBR scores considering mutations ≥ 5 (A) Scatterplot of logit-mutation probability versus log-PHBR. (B) GAM-estimated logit-mutation probability versus log-PHBR score. (C–F) ORs (black squares) and their 95% CIs (discontinuous lines) for acquiring a mutation displayed for all cancer types for (C) the within-residue model for mutations occurring ≥ 5 times in TCGA and for (D) the within-patient model for mutations occurring ≥ 5 times in TCGA (E) within-residue model for mutations occurring ≥ 20 times in TCGA and (F) within-patient model for mutations occurring ≥ 20 times in TCGA. (G) A ROC curve showing the accuracy of the PHBR and the PBR for classifying the extracellular presentation of a residue by a patient’s six MHC alleles. The aggregated PHBR/PBR presentation scores for 5 cell lines expressing 6 MHC alleles was compared to the PHBR/PBR scores for a random set of residues based on the same MHC alleles. (D) Error bars denote the 1.5 IQR range. When we analyzed the relationship between log-PHBR and the logit mutation probability within patients, we found that the log-PHBR was positively associated with a significant increase in the odds of a patient acquiring a mutation, supporting that patients have a higher probability of acquiring mutations less effectively presented by their MHC-I (within-patient model; mutation frequency ≥5; odds ratio [OR] = 1.28; 95% confidence interval [CI] [1.25, 1.31]; p < 2e-16) (Table 1). For each unit increase in log-PHBR, the odds of a mutation increases by 28%. The influence of PHBR tended to be stronger for mutations that were observed more frequently, for example, for mutations observed at least 20 times each unit increase in log-PHBR resulted in an odds increase of 54.5% (Table S4). These results demonstrate that the PHBR score is predictive of which recurrent oncogenic mutations are likely to drive an individual’s tumor during the early stages of tumor development pan-cancer.Table 1Quantitative Estimate of the Association between PHBR Score and Mutation OccurrenceWithin ResidueWithin PatientOR95% CIp ValueOR95% CIp Value≥5 mutations1.030.997,1.060.171.281.25, 1.31<2 × 10−16Passenger mutations10.97,1.030.9510.96, 1.030.97Germline variants0.9970.994,0.9990.150.9950.993, 0.9965.8 × 10−10ORs, 95% CIs, and p values are shown for within-mutation and within-patient models relating PHBR score to mutations observed ≥5 times across tumors. Models relating PHBR score to a set of 1,000 passenger mutations and 1,000 germline variants serve as controls. See also Table S4. Open table in a new tab ORs, 95% CIs, and p values are shown for within-mutation and within-patient models relating PHBR score to mutations observed ≥5 times across tumors. Models relating PHBR score to a set of 1,000 passenger mutations and 1,000 germline variants serve as controls. See also Table S4. The second approach analyzes the relationship between the log-PHBR and logit mutation probability across patients with the same mutation. This formulation evaluates whether the log-PHBR of a mutation has predictive power to determine which patients are at higher risk when the probability of the mutation occurring among patients is already known. This analysis revealed that PHBR was not a significant predictor of which patients would obtain a specific mutation (within-mutation model; mutation frequency ≥5; OR = 1.03; 95% CI [0.99, 1.06]; p < 0.17). The negligible effect of log-PHBR in this setting suggests that its influence is generally already captured by the random effect that models variation in mutation frequency and that incorporates the PHBR influence demonstrated by the within patients analysis. To determine whether PHBR score predictive power is distributed equally among different cancer types, we repeated the analysis within groups of at least 100 patients with a common tumor type (Figures 4D and 4E; all tumor types shown in Figures S4C and S4D). Once again, the within-mutation analysis mostly returned ORs with 95% CIs that included 1, indicating a lack of predictive power (Figure 4D, Table S5A). In clear contrast, the within-patient analysis returned multiple tumor types for which the OR was significantly greater than 1 (Figure 4E; Table S5B). While PHBR was predictive of mutation occurrence within-patient in more than 50% of the tumor types evaluated, there were clear differences in the magnitudes of the ORs, suggesting that PHBR could be more predictive in some tumor types than others. The strongest effects were observed in thyroid cancer (OR = 2.51; 95% CI [2.25, 2.8]), while no association was observed in acute myeloid leukemia, lung squamous cell carcinoma, sarcoma, or clear cell renal carcinoma. Notably, effect sizes were even larger when we considered only mutations observed >20 times across tumors (Figures S4E and S4F; T" @default.
• W4237687615 created "2022-05-12" @default.
• W4237687615 creator A5006783362 @default.
• W4237687615 creator A5007846743 @default.
• W4237687615 creator A5009470472 @default.
• W4237687615 creator A5027615447 @default.
• W4237687615 creator A5049102885 @default.
• W4237687615 creator A5055754716 @default.
• W4237687615 creator A5063900043 @default.
• W4237687615 creator A5064193594 @default.
• W4237687615 creator A5080116477 @default.
• W4237687615 creator A5086091950 @default.
• W4237687615 creator A5089373684 @default.
• W4237687615 date "2017-11-01" @default.
• W4237687615 modified "2023-10-14" @default.
• W4237687615 title "MHC-I Genotype Restricts the Oncogenic Mutational Landscape" @default.
• W4237687615 cites W1642873092 @default.
• W4237687615 cites W1964154017 @default.
• W4237687615 cites W1977385060 @default.
• W4237687615 cites W1987908519 @default.
• W4237687615 cites W2007026865 @default.
• W4237687615 cites W2009483826 @default.
• W4237687615 cites W2026318599 @default.
• W4237687615 cites W2028433016 @default.
• W4237687615 cites W2049553585 @default.
• W4237687615 cites W2053707749 @default.
• W4237687615 cites W2062889582 @default.
• W4237687615 cites W2097995306 @default.
• W4237687615 cites W2098951833 @default.
• W4237687615 cites W2101653483 @default.
• W4237687615 cites W2107079238 @default.
• W4237687615 cites W2112752664 @default.
• W4237687615 cites W2117692326 @default.
• W4237687615 cites W2118151336 @default.
• W4237687615 cites W2119160928 @default.
• W4237687615 cites W2134182833 @default.
• W4237687615 cites W2135078307 @default.
• W4237687615 cites W2140968934 @default.
• W4237687615 cites W2141609931 @default.
• W4237687615 cites W2150056019 @default.
• W4237687615 cites W2154974159 @default.
• W4237687615 cites W2155786171 @default.
• W4237687615 cites W2160834915 @default.
• W4237687615 cites W2172170289 @default.
• W4237687615 cites W2214699570 @default.
• W4237687615 cites W2216333908 @default.
• W4237687615 cites W2267653167 @default.
• W4237687615 cites W2326870000 @default.
• W4237687615 cites W2406591275 @default.
• W4237687615 cites W2560367415 @default.
• W4237687615 cites W2589139221 @default.
• W4237687615 cites W2600132724 @default.
• W4237687615 cites W2739999456 @default.
• W4237687615 cites W4210659358 @default.
• W4237687615 cites W4252232449 @default.
• W4237687615 doi "https://doi.org/10.1016/j.cell.2017.09.050" @default.
• W4237687615 hasPubMedId "https://pubmed.ncbi.nlm.nih.gov/29107334" @default.
• W4237687615 hasPublicationYear "2017" @default.
• W4237687615 type Work @default.
• W4237687615 citedByCount "246" @default.
• W4237687615 countsByYear W42376876152017 @default.
• W4237687615 countsByYear W42376876152018 @default.
• W4237687615 countsByYear W42376876152019 @default.
• W4237687615 countsByYear W42376876152020 @default.
• W4237687615 countsByYear W42376876152021 @default.
• W4237687615 countsByYear W42376876152022 @default.
• W4237687615 countsByYear W42376876152023 @default.
• W4237687615 crossrefType "journal-article" @default.
• W4237687615 hasAuthorship W4237687615A5006783362 @default.
• W4237687615 hasAuthorship W4237687615A5007846743 @default.
• W4237687615 hasAuthorship W4237687615A5009470472 @default.
• W4237687615 hasAuthorship W4237687615A5027615447 @default.
• W4237687615 hasAuthorship W4237687615A5049102885 @default.
• W4237687615 hasAuthorship W4237687615A5055754716 @default.
• W4237687615 hasAuthorship W4237687615A5063900043 @default.
• W4237687615 hasAuthorship W4237687615A5064193594 @default.
• W4237687615 hasAuthorship W4237687615A5080116477 @default.
• W4237687615 hasAuthorship W4237687615A5086091950 @default.
• W4237687615 hasAuthorship W4237687615A5089373684 @default.
• W4237687615 hasBestOaLocation W42376876151 @default.
• W4237687615 hasConcept C104317684 @default.
• W4237687615 hasConcept C135763542 @default.
• W4237687615 hasConcept C207936829 @default.
• W4237687615 hasConcept C501734568 @default.
• W4237687615 hasConcept C54355233 @default.
• W4237687615 hasConcept C78458016 @default.
• W4237687615 hasConcept C86803240 @default.
• W4237687615 hasConceptScore W4237687615C104317684 @default.
• W4237687615 hasConceptScore W4237687615C135763542 @default.
• W4237687615 hasConceptScore W4237687615C207936829 @default.
• W4237687615 hasConceptScore W4237687615C501734568 @default.
• W4237687615 hasConceptScore W4237687615C54355233 @default.
• W4237687615 hasConceptScore W4237687615C78458016 @default.
• W4237687615 hasConceptScore W4237687615C86803240 @default.
• W4237687615 hasFunder F4320332161 @default.
• W4237687615 hasIssue "6" @default.
• W4237687615 hasLocation W42376876151 @default.
• W4237687615 hasLocation W42376876152 @default.

• next