Fine-mapping analysis including over 254,000 East Asian and European descendants identifies 136 putative colorectal cancer susceptibility genes.

Genome-wide association studies (GWAS) have identified more than 200 common genetic variants independently associated with colorectal cancer (CRC) risk, but the causal variants and target genes are mostly unknown. We sought to fine-map all known CRC risk loci using GWAS data from 100,204 cases and 154,587 controls of East Asian and European ancestry. Our stepwise conditional analyses revealed 238 independent association signals of CRC risk, each with a set of credible causal variants (CCVs), of which 28 signals had a single CCV. Our cis-eQTL/mQTL and colocalization analyses using colorectal tissue-specific transcriptome and methylome data separately from 1299 and 321 individuals, along with functional genomic investigation, uncovered 136 putative CRC susceptibility genes, including 56 genes not previously reported. Analyses of single-cell RNA-seq data from colorectal tissues revealed 17 putative CRC susceptibility genes with distinct expression patterns in specific cell types. Analyses of whole exome sequencing data provided additional support for several target genes identified in this study as CRC susceptibility genes. Enrichment analyses of the 136 genes uncover pathways not previously linked to CRC risk. Our study substantially expanded association signals for CRC and provided additional insight into the biological mechanisms underlying CRC development.

Novel insights into genetic susceptibility for colorectal cancer from transcriptome-wide association and functional investigation.

BACKGROUND: Transcriptome-wide association studies have been successful in identifying candidate susceptibility genes for colorectal cancer (CRC). To strengthen susceptibility gene discovery, we conducted a large transcriptome-wide association study and an alternative splicing transcriptome-wide association study in CRC using improved genetic prediction models and performed in-depth functional investigations. METHODS: We analyzed RNA-sequencing data from normal colon tissues and genotype data from 423 European descendants to build genetic prediction models of gene expression and alternative splicing and evaluated model performance using independent RNA-sequencing data from normal colon tissues of the Genotype-Tissue Expression Project. We applied the verified models to genome-wide association studies (GWAS) summary statistics among 58 131 CRC cases and 67 347 controls of European ancestry to evaluate associations of genetically predicted gene expression and alternative splicing with CRC risk. We performed in vitro functional assays for 3 selected genes in multiple CRC cell lines. RESULTS: We identified 57 putative CRC susceptibility genes, which included the 48 genes from transcriptome-wide association studies and 15 genes from splicing transcriptome-wide association studies, at a Bonferroni-corrected P value less than .05. Of these, 16 genes were not previously implicated in CRC susceptibility, including a gene PDE7B (6q23.3) at locus previously not reported by CRC GWAS. Gene knockdown experiments confirmed the oncogenic roles for 2 unreported genes, TRPS1 and METRNL, and a recently reported gene, C14orf166. CONCLUSION: This study discovered new putative susceptibility genes of CRC and provided novel insights into the biological mechanisms underlying CRC development.

Comparing ancestry calibration approaches for a trans-ancestry colorectal cancer polygenic risk score.

Background: Colorectal cancer (CRC) is a complex disease with monogenic, polygenic and environmental risk factors. Polygenic risk scores (PRS) are being developed to identify high polygenic risk individuals. Due to differences in genetic background, PRS distributions vary by ancestry, necessitating calibration. Methods: We compared four calibration methods using the All of Us Research Program Whole Genome Sequence data for a CRC PRS previously developed in participants of European and East Asian ancestry. The methods contrasted results from linear models with A) the entire data set or an ancestrally diverse training set AND B) covariates including principal components of ancestry or admixture. Calibration with the training set adjusted the variance in addition to the mean. Results: All methods performed similarly within ancestry with OR (95% C.I.) per s.d. change in PRS: African 1.5 (1.02, 2.08), Admixed American 2.2 (1.27, 3.85), European 1.6 (1.43, 1.89), and Middle Eastern 1.1 (0.71, 1.63). Using admixture and an ancestrally diverse training set provided distributions closest to standard Normal with accurate upper tail frequencies. Conclusion: Although the PRS is predictive of CRC risk for most ancestries, its performance varies by ancestry. Post-hoc calibration preserves the risk prediction within ancestries. Training a calibration model on ancestrally diverse participants to adjust both the mean and variance of the PRS, using admixture as covariates, created standard Normal z-scores. These z-scores can be used to identify patients at high polygenic risk, and can be incorporated into comprehensive risk scores including other known risk factors, allowing for more precise risk estimates.

Combining Asian and European genome-wide association studies of colorectal cancer improves risk prediction across racial and ethnic populations.

Polygenic risk scores (PRS) have great potential to guide precision colorectal cancer (CRC) prevention by identifying those at higher risk to undertake targeted screening. However, current PRS using European ancestry data have sub-optimal performance in non-European ancestry populations, limiting their utility among these populations. Towards addressing this deficiency, we expand PRS development for CRC by incorporating Asian ancestry data (21,731 cases; 47,444 controls) into European ancestry training datasets (78,473 cases; 107,143 controls). The AUC estimates (95% CI) of PRS are 0.63(0.62-0.64), 0.59(0.57-0.61), 0.62(0.60-0.63), and 0.65(0.63-0.66) in independent datasets including 1681-3651 cases and 8696-115,105 controls of Asian, Black/African American, Latinx/Hispanic, and non-Hispanic White, respectively. They are significantly better than the European-centric PRS in all four major US racial and ethnic groups (p-values < 0.05). Further inclusion of non-European ancestry populations, especially Black/African American and Latinx/Hispanic, is needed to improve the risk prediction and enhance equity in applying PRS in clinical practice.

Risk-Stratified Screening for Colorectal Cancer Using Genetic and Environmental Risk Factors: A Cost-Effectiveness Analysis Based on Real-World Data.

BACKGROUND & AIMS: Previous studies on the cost-effectiveness of personalized colorectal cancer (CRC) screening were based on hypothetical performance of CRC risk prediction and did not consider the association with competing causes of death. In this study, we estimated the cost-effectiveness of risk-stratified screening using real-world data for CRC risk and competing causes of death. METHODS: Risk predictions for CRC and competing causes of death from a large community-based cohort were used to stratify individuals into risk groups. A microsimulation model was used to optimize colonoscopy screening for each risk group by varying the start age (40-60 years), end age (70-85 years), and screening interval (5-15 years). The outcomes included personalized screening ages and intervals and cost-effectiveness compared with uniform colonoscopy screening (ages 45-75, every 10 years). Key assumptions were varied in sensitivity analyses. RESULTS: Risk-stratified screening resulted in substantially different screening recommendations, ranging from a one-time colonoscopy at age 60 for low-risk individuals to a colonoscopy every 5 years from ages 40 to 85 for high-risk individuals. Nevertheless, on a population level, risk-stratified screening would increase net quality-adjusted life years gained (QALYG) by only 0.7% at equal costs to uniform screening or reduce average costs by 1.2% for equal QALYG. The benefit of risk-stratified screening improved when it was assumed to increase participation or costs less per genetic test. CONCLUSIONS: Personalized screening for CRC, accounting for competing causes of death risk, could result in highly tailored individual screening programs. However, average improvements across the population in QALYG and cost-effectiveness compared with uniform screening are small.

Combining Asian-European Genome-Wide Association Studies of Colorectal Cancer Improves Risk Prediction Across Race and Ethnicity.

Polygenic risk scores (PRS) have great potential to guide precision colorectal cancer (CRC) prevention by identifying those at higher risk to undertake targeted screening. However, current PRS using European ancestry data have sub-optimal performance in non-European ancestry populations, limiting their utility among these populations. Towards addressing this deficiency, we expanded PRS development for CRC by incorporating Asian ancestry data (21,731 cases; 47,444 controls) into European ancestry training datasets (78,473 cases; 107,143 controls). The AUC estimates (95% CI) of PRS were 0.63(0.62-0.64), 0.59(0.57-0.61), 0.62(0.60-0.63), and 0.65(0.63-0.66) in independent datasets including 1,681-3,651 cases and 8,696-115,105 controls of Asian, Black/African American, Latinx/Hispanic, and non-Hispanic White, respectively. They were significantly better than the European-centric PRS in all four major US racial and ethnic groups (p-values<0.05). Further inclusion of non-European ancestry populations, especially Black/African American and Latinx/Hispanic, is needed to improve the risk prediction and enhance equity in applying PRS in clinical practice.

Validation of a Genetic-Enhanced Risk Prediction Model for Colorectal Cancer in a Large Community-Based Cohort.

BACKGROUND: Polygenic risk scores (PRS) which summarize individuals' genetic risk profile may enhance targeted colorectal cancer screening. A critical step towards clinical implementation is rigorous external validations in large community-based cohorts. This study externally validated a PRS-enhanced colorectal cancer risk model comprising 140 known colorectal cancer loci to provide a comprehensive assessment on prediction performance. METHODS: The model was developed using 20,338 individuals and externally validated in a community-based cohort (n = 85,221). We validated predicted 5-year absolute colorectal cancer risk, including calibration using expected-to-observed case ratios (E/O) and calibration plots, and discriminatory accuracy using time-dependent AUC. The PRS-related improvement in AUC, sensitivity and specificity were assessed in individuals of age 45 to 74 years (screening-eligible age group) and 40 to 49 years with no endoscopy history (younger-age group). RESULTS: In European-ancestral individuals, the predicted 5-year risk calibrated well [E/O = 1.01; 95% confidence interval (CI), 0.91-1.13] and had high discriminatory accuracy (AUC = 0.73; 95% CI, 0.71-0.76). Adding the PRS to a model with age, sex, family and endoscopy history improved the 5-year AUC by 0.06 (P < 0.001) and 0.14 (P = 0.05) in the screening-eligible age and younger-age groups, respectively. Using a risk-threshold of 5-year SEER colorectal cancer incidence rate at age 50 years, adding the PRS had a similar sensitivity but improved the specificity by 11% (P < 0.001) in the screening-eligible age group. In the younger-age group it improved the sensitivity by 27% (P = 0.04) with similar specificity. CONCLUSIONS: The proposed PRS-enhanced model provides a well-calibrated 5-year colorectal cancer risk prediction and improves discriminatory accuracy in the external cohort. IMPACT: The proposed model has potential utility in risk-stratified colorectal cancer prevention.

Deciphering colorectal cancer genetics through multi-omic analysis of 100,204 cases and 154,587 controls of European and east Asian ancestries.

Colorectal cancer (CRC) is a leading cause of mortality worldwide. We conducted a genome-wide association study meta-analysis of 100,204 CRC cases and 154,587 controls of European and east Asian ancestry, identifying 205 independent risk associations, of which 50 were unreported. We performed integrative genomic, transcriptomic and methylomic analyses across large bowel mucosa and other tissues. Transcriptome- and methylome-wide association studies revealed an additional 53 risk associations. We identified 155 high-confidence effector genes functionally linked to CRC risk, many of which had no previously established role in CRC. These have multiple different functions and specifically indicate that variation in normal colorectal homeostasis, proliferation, cell adhesion, migration, immunity and microbial interactions determines CRC risk. Crosstissue analyses indicated that over a third of effector genes most probably act outside the colonic mucosa. Our findings provide insights into colorectal oncogenesis and highlight potential targets across tissues for new CRC treatment and chemoprevention strategies.

Risk Stratification for Early-Onset Colorectal Cancer Using a Combination of Genetic and Environmental Risk Scores: An International Multi-Center Study.

BACKGROUND: The incidence of colorectal cancer (CRC) among individuals aged younger than 50 years has been increasing. As screening guidelines lower the recommended age of screening initiation, concerns including the burden on screening capacity and costs have been recognized, suggesting that an individualized approach may be warranted. We developed risk prediction models for early-onset CRC that incorporate an environmental risk score (ERS), including 16 lifestyle and environmental factors, and a polygenic risk score (PRS) of 141 variants. METHODS: Relying on risk score weights for ERS and PRS derived from studies of CRC at all ages, we evaluated risks for early-onset CRC in 3486 cases and 3890 controls aged younger than 50 years. Relative and absolute risks for early-onset CRC were assessed according to values of the ERS and PRS. The discriminatory performance of these scores was estimated using the covariate-adjusted area under the receiver operating characteristic curve. RESULTS: Increasing values of ERS and PRS were associated with increasing relative risks for early-onset CRC (odds ratio per SD of ERS = 1.14, 95% confidence interval [CI] = 1.08 to 1.20; odds ratio per SD of PRS = 1.59, 95% CI = 1.51 to 1.68), both contributing to case-control discrimination (area under the curve = 0.631, 95% CI = 0.615 to 0.647). Based on absolute risks, we can expect 26 excess cases per 10 000 men and 21 per 10 000 women among those scoring at the 90th percentile for both risk scores. CONCLUSIONS: Personal risk scores have the potential to identify individuals at differential relative and absolute risk for early-onset CRC. Improved discrimination may aid in targeted CRC screening of younger, high-risk individuals, potentially improving outcomes.

Genome-wide Modeling of Polygenic Risk Score in Colorectal Cancer Risk.

Accurate colorectal cancer (CRC) risk prediction models are critical for identifying individuals at low and high risk of developing CRC, as they can then be offered targeted screening and interventions to address their risks of developing disease (if they are in a high-risk group) and avoid unnecessary screening and interventions (if they are in a low-risk group). As it is likely that thousands of genetic variants contribute to CRC risk, it is clinically important to investigate whether these genetic variants can be used jointly for CRC risk prediction. In this paper, we derived and compared different approaches to generating predictive polygenic risk scores (PRS) from genome-wide association studies (GWASs) including 55,105 CRC-affected case subjects and 65,079 control subjects of European ancestry. We built the PRS in three ways, using (1) 140 previously identified and validated CRC loci; (2) SNP selection based on linkage disequilibrium (LD) clumping followed by machine-learning approaches; and (3) LDpred, a Bayesian approach for genome-wide risk prediction. We tested the PRS in an independent cohort of 101,987 individuals with 1,699 CRC-affected case subjects. The discriminatory accuracy, calculated by the age- and sex-adjusted area under the receiver operating characteristics curve (AUC), was highest for the LDpred-derived PRS (AUC = 0.654) including nearly 1.2 M genetic variants (the proportion of causal genetic variants for CRC assumed to be 0.003), whereas the PRS of the 140 known variants identified from GWASs had the lowest AUC (AUC = 0.629). Based on the LDpred-derived PRS, we are able to identify 30% of individuals without a family history as having risk for CRC similar to those with a family history of CRC, whereas the PRS based on known GWAS variants identified only top 10% as having a similar relative risk. About 90% of these individuals have no family history and would have been considered average risk under current screening guidelines, but might benefit from earlier screening. The developed PRS offers a way for risk-stratified CRC screening and other targeted interventions.

Cumulative Burden of Colorectal Cancer-Associated Genetic Variants Is More Strongly Associated With Early-Onset vs Late-Onset Cancer.

BACKGROUND & AIMS: Early-onset colorectal cancer (CRC, in persons younger than 50 years old) is increasing in incidence; yet, in the absence of a family history of CRC, this population lacks harmonized recommendations for prevention. We aimed to determine whether a polygenic risk score (PRS) developed from 95 CRC-associated common genetic risk variants was associated with risk for early-onset CRC. METHODS: We studied risk for CRC associated with a weighted PRS in 12,197 participants younger than 50 years old vs 95,865 participants 50 years or older. PRS was calculated based on single nucleotide polymorphisms associated with CRC in a large-scale genome-wide association study as of January 2019. Participants were pooled from 3 large consortia that provided clinical and genotyping data: the Colon Cancer Family Registry, the Colorectal Transdisciplinary Study, and the Genetics and Epidemiology of Colorectal Cancer Consortium and were all of genetically defined European descent. Findings were replicated in an independent cohort of 72,573 participants. RESULTS: Overall associations with CRC per standard deviation of PRS were significant for early-onset cancer, and were stronger compared with late-onset cancer (P for interaction = .01); when we compared the highest PRS quartile with the lowest, risk increased 3.7-fold for early-onset CRC (95% CI 3.28-4.24) vs 2.9-fold for late-onset CRC (95% CI 2.80-3.04). This association was strongest for participants without a first-degree family history of CRC (P for interaction = 5.61 × 10-5). When we compared the highest with the lowest quartiles in this group, risk increased 4.3-fold for early-onset CRC (95% CI 3.61-5.01) vs 2.9-fold for late-onset CRC (95% CI 2.70-3.00). Sensitivity analyses were consistent with these findings. CONCLUSIONS: In an analysis of associations with CRC per standard deviation of PRS, we found the cumulative burden of CRC-associated common genetic variants to associate with early-onset cancer, and to be more strongly associated with early-onset than late-onset cancer, particularly in the absence of CRC family history. Analyses of PRS, along with environmental and lifestyle risk factors, might identify younger individuals who would benefit from preventive measures.

New bandwidth selection criterion for Kernel PCA: approach to dimensionality reduction and classification problems.

BACKGROUND: DNA microarrays are potentially powerful technology for improving diagnostic classification, treatment selection, and prognostic assessment. The use of this technology to predict cancer outcome has a history of almost a decade. Disease class predictors can be designed for known disease cases and provide diagnostic confirmation or clarify abnormal cases. The main input to this class predictors are high dimensional data with many variables and few observations. Dimensionality reduction of these features set significantly speeds up the prediction task. Feature selection and feature transformation methods are well known preprocessing steps in the field of bioinformatics. Several prediction tools are available based on these techniques. RESULTS: Studies show that a well tuned Kernel PCA (KPCA) is an efficient preprocessing step for dimensionality reduction, but the available bandwidth selection method for KPCA was computationally expensive. In this paper, we propose a new data-driven bandwidth selection criterion for KPCA, which is related to least squares cross-validation for kernel density estimation. We propose a new prediction model with a well tuned KPCA and Least Squares Support Vector Machine (LS-SVM). We estimate the accuracy of the newly proposed model based on 9 case studies. Then, we compare its performances (in terms of test set Area Under the ROC Curve (AUC) and computational time) with other well known techniques such as whole data set + LS-SVM, PCA + LS-SVM, t-test + LS-SVM, Prediction Analysis of Microarrays (PAM) and Least Absolute Shrinkage and Selection Operator (Lasso). Finally, we assess the performance of the proposed strategy with an existing KPCA parameter tuning algorithm by means of two additional case studies. CONCLUSION: We propose, evaluate, and compare several mathematical/statistical techniques, which apply feature transformation/selection for subsequent classification, and consider its application in medical diagnostics. Both feature selection and feature transformation perform well on classification tasks. Due to the dynamic selection property of feature selection, it is hard to define significant features for the classifier, which predicts classes of future samples. Moreover, the proposed strategy enjoys a distinctive advantage with its relatively lesser time complexity.

Predicting breast cancer using an expression values weighted clinical classifier.

BACKGROUND: Clinical data, such as patient history, laboratory analysis, ultrasound parameters-which are the basis of day-to-day clinical decision support-are often used to guide the clinical management of cancer in the presence of microarray data. Several data fusion techniques are available to integrate genomics or proteomics data, but only a few studies have created a single prediction model using both gene expression and clinical data. These studies often remain inconclusive regarding an obtained improvement in prediction performance. To improve clinical management, these data should be fully exploited. This requires efficient algorithms to integrate these data sets and design a final classifier. LS-SVM classifiers and generalized eigenvalue/singular value decompositions are successfully used in many bioinformatics applications for prediction tasks. While bringing up the benefits of these two techniques, we propose a machine learning approach, a weighted LS-SVM classifier to integrate two data sources: microarray and clinical parameters. RESULTS: We compared and evaluated the proposed methods on five breast cancer case studies. Compared to LS-SVM classifier on individual data sets, generalized eigenvalue decomposition (GEVD) and kernel GEVD, the proposed weighted LS-SVM classifier offers good prediction performance, in terms of test area under ROC Curve (AUC), on all breast cancer case studies. CONCLUSIONS: Thus a clinical classifier weighted with microarray data set results in significantly improved diagnosis, prognosis and prediction responses to therapy. The proposed model has been shown as a promising mathematical framework in both data fusion and non-linear classification problems.

Maximum Likelihood Estimation of GEVD: Applications in Bioinformatics.

We propose a method, maximum likelihood estimation of generalized eigenvalue decomposition (MLGEVD) that employs a well known technique relying on the generalization of singular value decomposition (SVD). The main aim of the work is to show the tight equivalence between MLGEVD and generalized ridge regression. This relationship reveals an important mathematical property of GEVD in which the second argument act as prior information in the model. Thus we show that MLGEVD allows the incorporation of external knowledge about the quantities of interest into the estimation problem. We illustrate the importance of prior knowledge in clinical decision making/identifying differentially expressed genes with case studies for which microarray data sets with corresponding clinical/literature information are available. On all of these three case studies, MLGEVD outperformed GEVD on prediction in terms of test area under the ROC curve (test AUC). MLGEVD results in significantly improved diagnosis, prognosis and prediction of therapy response.