Cardiometabolic phenotypes and mitochondrial DNA copy number in two cohorts of UK women.

The mitochondrial genome is present at variable copy number between individuals. Mitochondria are vulnerable to oxidative stress, and their dysfunction may be associated with cardiovascular disease. The association of mitochondrial DNA copy number with cardiometabolic risk factors (lipids, glycaemic traits, inflammatory markers, anthropometry and blood pressure) was assessed in two independent cohorts of European origin women, one in whom outcomes were measured at mean (SD) age 30 (4.3) years (N=2278) and the second at 69.4 (5.5) years (N=2872). Mitochondrial DNA copy number was assayed by quantitative polymerase chain reaction. Associations were adjusted for smoking, sociodemographic status, laboratory factors and white cell traits. Out of a total of 12 outcomes assessed in both cohorts, mitochondrial DNA copy number showed little or no association with the majority (point estimates were close to zero and nearly all p-values were >0.01). The strongest evidence was for an inverse association in the older cohort with insulin (standardised beta [95%CI]: -0.06, [-0.098, -0.022], p=0.002), but this association did not replicate in the younger cohort. Our findings do not provide support for variation in mitochondrial DNA copy number having an important impact on cardio-metabolic risk factors in European origin women.

Possible positive effect of the APOE ε2 allele on cognition in early to mid-adult life.

BACKGROUND: ε4 allele possession is associated with an increased risk of Alzheimer's disease. Its effects earlier in life are less well understood. Previous studies have reported both detrimental effects and a lack of effect on cognition outside dementia. We used genotype based recall from the ALSPAC study to investigate whether APOE genotype influences cognition in earlier adult life. METHODS: We invited all individuals with the rarer ε22 or ε44 genotypes and equal numbers of those with ε32, ε33 or ε34 APOE genotypes (total n invited = 1936, ages 23-67). Participants were screened for dementia using the Addenbrooke's Cognitive Examination Revised (ACE-R). Participants were asked to complete a 3 h battery of neuropsychological tests covering a range of cognitive domains. The primary outcome was performance on the Rey Auditory Verbal Learning Test (RAVLT). Transformation of variables was used where required to permit parametric testing. As genotypes are unlikely to be confounded unadjusted analyses were performed. RESULTS: 114 participants were recruited to the study (39 ε33, 27 ε34, 15 ε44, 26 ε32 & 7 ε22). ε4+ participants had higher scores on the cognitive failures questionnaire (10 point increase, p = 0.006) but no deficits on objective cognitive testing. ε2 carriers had slightly better episodic memory performance (p = 0.016), slightly improved n-back accuracy and better executive functioning (trails A&B, p = 0.005). CONCLUSIONS: It is intriguing that the ε2+ group performed better as this group have a lower risk of Alzheimer's disease. Most previous studies have analysed as ε4/non ε4 so may have missed this effect.

HAPRAP: a haplotype-based iterative method for statistical fine mapping using GWAS summary statistics.

MOTIVATION: Fine mapping is a widely used approach for identifying the causal variant(s) at disease-associated loci. Standard methods (e.g. multiple regression) require individual level genotypes. Recent fine mapping methods using summary-level data require the pairwise correlation coefficients ([Formula: see text]) of the variants. However, haplotypes rather than pairwise [Formula: see text], are the true biological representation of linkage disequilibrium (LD) among multiple loci. In this article, we present an empirical iterative method, HAPlotype Regional Association analysis Program (HAPRAP), that enables fine mapping using summary statistics and haplotype information from an individual-level reference panel. RESULTS: Simulations with individual-level genotypes show that the results of HAPRAP and multiple regression are highly consistent. In simulation with summary-level data, we demonstrate that HAPRAP is less sensitive to poor LD estimates. In a parametric simulation using Genetic Investigation of ANthropometric Traits height data, HAPRAP performs well with a small training sample size (N < 2000) while other methods become suboptimal. Moreover, HAPRAP's performance is not affected substantially by single nucleotide polymorphisms (SNPs) with low minor allele frequencies. We applied the method to existing quantitative trait and binary outcome meta-analyses (human height, QTc interval and gallbladder disease); all previous reported association signals were replicated and two additional variants were independently associated with human height. Due to the growing availability of summary level data, the value of HAPRAP is likely to increase markedly for future analyses (e.g. functional prediction and identification of instruments for Mendelian randomization). AVAILABILITY AND IMPLEMENTATION: The HAPRAP package and documentation are available at http://apps.biocompute.org.uk/haprap/ CONTACT: : jie.zheng@bristol.ac.uk or tom.gaunt@bristol.ac.ukSupplementary information: Supplementary data are available at Bioinformatics online.

Importance of Genetic Studies in Consanguineous Populations for the Characterization of Novel Human Gene Functions.

Consanguineous offspring have elevated levels of homozygosity. Autozygous stretches within their genome are likely to harbour loss of function (LoF) mutations which will lead to complete inactivation or dysfunction of genes. Studying consanguineous offspring with clinical phenotypes has been very useful for identifying disease causal mutations. However, at present, most of the genes in the human genome have no disorder associated with them or have unknown function. This is presumably mostly due to the fact that homozygous LoF variants are not observed in outbred populations which are the main focus of large sequencing projects. However, another reason may be that many genes in the genome-even when completely "knocked out," do not cause a distinct or defined phenotype. Here, we discuss the benefits and implications of studying consanguineous populations, as opposed to the traditional approach of analysing a subset of consanguineous families or individuals with disease. We suggest that studying consanguineous populations "as a whole" can speed up the characterisation of novel gene functions as well as indicating nonessential genes and/or regions in the human genome. We also suggest designing a single nucleotide variant (SNV) array to make the process more efficient.

A study of common Mendelian disease carriers across ageing British cohorts: meta-analyses reveal heterozygosity for alpha 1-antitrypsin deficiency increases respiratory capacity and height.

BACKGROUND: Several recessive Mendelian disorders are common in Europeans, including cystic fibrosis (CFTR), medium-chain-acyl-Co-A-dehydrogenase deficiency (ACADM), phenylketonuria (PAH) and alpha 1-antitrypsin deficiency (SERPINA1). METHODS: In a multicohort study of >19,000 older individuals, we investigated the relevant phenotypes in heterozygotes for these genes: lung function (forced expiratory volume in 1 second (FEV1), forced vital capacity (FVC)) for CFTR and SERPINA1; cognitive measures for ACADM and PAH; and physical capability for ACADM, PAH and SERPINA1. RESULTS: Findings were mostly negative but lung function in SERPINA1 (protease inhibitor (PI) Z allele, rs28929474) showed enhanced FEV1 and FVC (0.13 z-score increase in FEV1 (p=1.7 × 10(-5)) and 0.16 z-score increase in FVC (p=5.2 × 10(-8))) in PI-MZ individuals. Height adjustment (a known, strong correlate of FEV1 and FVC) revealed strong positive height associations of the Z allele (1.50 cm increase in height (p=3.6 × 10(-10))). CONCLUSIONS: The PI-MZ rare (2%) SNP effect is nearly four times greater than the 'top' common height SNP in HMGA2. However, height only partially attenuates the SERPINA1-FEV1 or FVC association (around 50%) and vice versa. Height SNP variants have recently been shown to be positively selected collectively in North versus South Europeans, while the Z allele high frequency is localised to North Europe. Although PI-ZZ is clinically disadvantageous to lung function, PI-MZ increases both height and respiratory function; potentially a balanced polymorphism. Partial blockade of PI could conceivably form part of a future poly-therapeutic approach in very short children. The notion that elastase inhibition should benefit patients with chronic obstructive pulmonary disease may also merit re-evaluation. PI is already a therapeutic target: our findings invite a reconsideration of the optimum level in respiratory care and novel pathway potential for development of agents for the management of growth disorders.

Plasma urate concentration and risk of coronary heart disease: a Mendelian randomisation analysis.

BACKGROUND: Increased circulating plasma urate concentration is associated with an increased risk of coronary heart disease, but the extent of any causative effect of urate on risk of coronary heart disease is still unclear. In this study, we aimed to clarify any causal role of urate on coronary heart disease risk using Mendelian randomisation analysis. METHODS: We first did a fixed-effects meta-analysis of the observational association of plasma urate and risk of coronary heart disease. We then used a conventional Mendelian randomisation approach to investigate the causal relevance using a genetic instrument based on 31 urate-associated single nucleotide polymorphisms (SNPs). To account for potential pleiotropic associations of certain SNPs with risk factors other than urate, we additionally did both a multivariable Mendelian randomisation analysis, in which the genetic associations of SNPs with systolic and diastolic blood pressure, HDL cholesterol, and triglycerides were included as covariates, and an Egger Mendelian randomisation (MR-Egger) analysis to estimate a causal effect accounting for unmeasured pleiotropy. FINDINGS: In the meta-analysis of 17 prospective observational studies (166 486 individuals; 9784 coronary heart disease events) a 1 SD higher urate concentration was associated with an odds ratio (OR) for coronary heart disease of 1·07 (95% CI 1·04-1·10). The corresponding OR estimates from the conventional, multivariable adjusted, and Egger Mendelian randomisation analysis (58 studies; 198 598 individuals; 65 877 events) were 1·18 (95% CI 1·08-1·29), 1·10 (1·00-1·22), and 1·05 (0·92-1·20), respectively, per 1 SD increment in plasma urate. INTERPRETATION: Conventional and multivariate Mendelian randomisation analysis implicates a causal role for urate in the development of coronary heart disease, but these estimates might be inflated by hidden pleiotropy. Egger Mendelian randomisation analysis, which accounts for pleiotropy but has less statistical power, suggests there might be no causal effect. These results might help investigators to determine the priority of trials of urate lowering for the prevention of coronary heart disease compared with other potential interventions. FUNDING: UK National Institute for Health Research, British Heart Foundation, and UK Medical Research Council.

Lipids, obesity and gallbladder disease in women: insights from genetic studies using the cardiovascular gene-centric 50K SNP array.

Gallbladder disease (GBD) has an overall prevalence of 10-40% depending on factors such as age, gender, population, obesity and diabetes, and represents a major economic burden. Although gallstones are composed of cholesterol by-products and are associated with obesity, presumed causal pathways remain unproven, although BMI reduction is typically recommended. We performed genetic studies to discover candidate genes and define pathways involved in GBD. We genotyped 15,241 women of European ancestry from three cohorts, including 3216 with GBD, using the Human cardiovascular disease (HumanCVD) BeadChip containing up to ~ 53,000 single-nucleotide polymorphisms (SNPs). Effect sizes with P-values for development of GBD were generated. We identify two new loci associated with GBD, GCKR rs1260326:T>C (P = 5.88 × 10(-7), ß = -0.146) and TTC39B rs686030:C>A (P = 6.95 x 10(-7), ß = 0.271) and detect four independent SNP effects in ABCG8 rs4953023:G>A (P=7.41 × 10(-47), ß = 0.734), ABCG8 rs4299376:G(>)T (P = 2.40 × 10(-18), ß = 0.278), ABCG5 rs6544718:T>C (P = 2.08 × 10(-14), ß = 0.044) and ABCG5 rs6720173:G>C (P = 3.81 × 10(-12), ß(=)0.262) in conditional analyses taking genotypes of rs4953023:G>A as a covariate. We also delineate the risk effects among many genotypes known to influence lipids. These data, from the largest GBD genetic study to date, show that specific, mainly hepatocyte-centred, components of lipid metabolism are important to GBD risk in women. We discuss the potential pharmaceutical implications of our findings.

Effect of smoking on physical and cognitive capability in later life: a multicohort study using observational and genetic approaches.

OBJECTIVES: The observed associations between smoking and functional measures at older ages are vulnerable to bias and confounding. Mendelian randomisation (MR) uses genotype as an instrumental variable to estimate unconfounded causal associations. We conducted a meta-analysis of the observational associations and implemented an MR approach using the smoking-related single nucleotide polymorphism rs16969968 to explore their causal nature. SETTING: 9 British cohorts belonging to the HALCyon collaboration. PARTICIPANTS: Individual participant data on N=26,692 individuals of European ancestry (N from earliest phase analysed per study) of mean ages 50-79 years were available for inclusion in observational meta-analyses of the primary outcomes. PRIMARY OUTCOMES: Physical capability, cognitive capability and cognitive decline. The smoking exposures were cigarettes per day, current versus ex-smoker, current versus never smoker and ever versus never smoker. RESULTS: In observational analyses current and ever smoking were generally associated with poorer physical and cognitive capability. For example, current smokers had a general fluid cognition score which was 0.17 z-score units (95% CI -0.221 to -0.124) lower than ex-smokers in cross-sectional analyses. Current smokers had a walk speed which was 0.25 z-score units lower than never smokers (95% CI -0.338 to -0.170). An MR instrumental variable approach for current versus ex-smoker and number of cigarettes smoked per day produced CIs which neither confirmed nor refuted the observational estimates. The number of genetic associations stratified by smoking status were consistent with type I error. CONCLUSIONS: Our observational analysis supports the hypothesis that smoking is detrimental to physical and cognitive capability. Further studies are needed for a suitably powered MR approach.

Possible Association of APOE Genotype with Working Memory in Young Adults.

BACKGROUND: Possession of the ε4 allele of the Apolipoprotein E (APOE) gene is associated with an increased risk of Alzheimer's disease. Early adult life effects of ε4 are less well understood. Working memory has been relatively little studied (compared to episodic memory) in relation to APOE genotype despite its importance in cognitive functioning. Our hypothesis was that ε4 would lead to an impairment in working memory in young adults. METHODS: We studied working memory using a computerised n-back task in the Avon Longitudinal Study of Parents and Children (ALSPAC) at age 18. Data was available for 1049-1927 participants and for the 2- and 3-back versions of the task. Using multiple and multi-level regression controlling for important confounders we examined the association between APOE genotype on accuracy and reaction times. RESULTS: There was no evidence of a genotype effect on accuracy when the two difficulty levels were examined separately. There was some evidence to support a deleterious effect of the ε4 allele on n-back accuracy in the multi-level regression. There was weak evidence that the ε22 group were less accurate but the numbers were very low in this group. The ε34 group had faster reaction times than the reference ε33 group in all adjusted analyses but the ε44 group were only faster in the 3-back condition in multi-level analyses. CONCLUSIONS: There was no evidence of benefit in ε4 carriers, but there was some evidence of a detrimental effect on working memory in this large study.

Identifying Highly Penetrant Disease Causal Mutations Using Next Generation Sequencing: Guide to Whole Process.

Recent technological advances have created challenges for geneticists and a need to adapt to a wide range of new bioinformatics tools and an expanding wealth of publicly available data (e.g., mutation databases, and software). This wide range of methods and a diversity of file formats used in sequence analysis is a significant issue, with a considerable amount of time spent before anyone can even attempt to analyse the genetic basis of human disorders. Another point to consider that is although many possess "just enough" knowledge to analyse their data, they do not make full use of the tools and databases that are available and also do not fully understand how their data was created. The primary aim of this review is to document some of the key approaches and provide an analysis schema to make the analysis process more efficient and reliable in the context of discovering highly penetrant causal mutations/genes. This review will also compare the methods used to identify highly penetrant variants when data is obtained from consanguineous individuals as opposed to nonconsanguineous; and when Mendelian disorders are analysed as opposed to common-complex disorders.

Proxy molecular diagnosis from whole-exome sequencing reveals Papillon-Lefevre syndrome caused by a missense mutation in CTSC.

Papillon-Lefevre syndrome (PLS) is an autosomal recessive disorder characterised by severe early onset periodontitis and palmoplantar hyperkeratosis. A previously reported missense mutation in the CTSC gene (NM_001814.4:c.899G>A:p.(G300D)) was identified in a homozygous state in two siblings diagnosed with PLS in a consanguineous family of Arabic ancestry. The variant was initially identified in a heterozygous state in a PLS unaffected sibling whose whole exome had been sequenced as part of a previous Primary ciliary dyskinesia study. Using this information, a proxy molecular diagnosis was made on the PLS affected siblings after consent was given to study this second disorder found to be segregating within the family. The prevalence of the mutation was then assayed in the local population using a representative sample of 256 unrelated individuals. The variant was absent in all subjects indicating that the variant is rare in Saudi Arabia. This family study illustrates how whole-exome sequencing can generate findings and inferences beyond its primary goal.

An integrative approach to predicting the functional effects of non-coding and coding sequence variation.

MOTIVATION: Technological advances have enabled the identification of an increasingly large spectrum of single nucleotide variants within the human genome, many of which may be associated with monogenic disease or complex traits. Here, we propose an integrative approach, named FATHMM-MKL, to predict the functional consequences of both coding and non-coding sequence variants. Our method utilizes various genomic annotations, which have recently become available, and learns to weight the significance of each component annotation source. RESULTS: We show that our method outperforms current state-of-the-art algorithms, CADD and GWAVA, when predicting the functional consequences of non-coding variants. In addition, FATHMM-MKL is comparable to the best of these algorithms when predicting the impact of coding variants. The method includes a confidence measure to rank order predictions.

Haptoglobin duplicon, hemoglobin, and vitamin C: analyses in the british women's heart and health study and Caerphilly prospective study.

BACKGROUND: Haptoglobin acts as an antioxidant by limiting peroxidative tissue damage by free hemoglobin. The haptoglobin gene allele Hp2 comprises a 1.7 kb partial duplication. Relative to allele Hp1, Hp2 carriers form protein multimers, suboptimal for hemoglobin scavenging. OBJECTIVE: To examine the association of haptoglobin genotype with a range of phenotypes, with emphasis on vitamin C and hemoglobin levels. METHODS: We applied a quantitative PCR assay for the duplication junction to two population cohorts including 2747 British women and 1198 British men. We examined the association of haptoglobin duplicon copy number with hemoglobin and vitamin C and used the copy number to complete a phenome scan. RESULTS: Hemoglobin concentrations were greater in those with Hp2,2 genotype, in women only (Hp1,1 13.45 g/dL, Hp1,2 13.49 g/dL, Hp2,2 13.61 g/dL; P = 0.002), though statistically there was no evidence of a difference between the sexes (z value = 1.2, P = 0.24). Haptoglobin genotype was not associated with vitamin C or any other phenotype in either cohort. CONCLUSIONS: Our results do not support association of haptoglobin genotype with vitamin C or with other phenotypes measured in two population cohorts. The apparent association between haptoglobin genotype and hemoglobin in the women's cohort merits further investigation.

Influence of adiposity-related genetic markers in a population of saudi arabians where other variables influencing obesity may be reduced.

Large scale studies in Europeans have clearly identified common polymorphism affecting BMI and obesity. We undertook a genotype study to examine the impact of variants, known to influence obesity, in a sample from the Saudi Arabian population, notable for its profound combination of low mean physical activity indices and high energy intake. Anthropometry measures and genotypes were obtained for 367 Saudis, taken from King Saud University and Biomarker Screening Project in Riyadh (Riyadh Cohort). We observed large effect sizes with obesity for rs10767664 (BDNF) (OR = 1.923, P = 0.00072) and rs3751812 (FTO) (OR = 1.523, P = 0.016) in our sample and, using weighted genetic risk scores, we found strong evidence of a cumulative effect using 11 SNPs taken predominantly from loci principally affecting appetite (OR = 2.57, P = 0.00092). We used conditional analyses to discern which of our three highly correlated FTO SNPs were responsible for the observed signal, although we were unable to determine with confidence which best marked the causal site. Our analysis indicates that markers located in loci known to influence fat mass through increased appetite affect obesity in Saudi Arabians to an extent possibly greater than in Europeans. Larger scale studies will be necessary to obtain a precise comparison.

Canonical correlation analysis for gene-based pleiotropy discovery.

Genome-wide association studies have identified a wealth of genetic variants involved in complex traits and multifactorial diseases. There is now considerable interest in testing variants for association with multiple phenotypes (pleiotropy) and for testing multiple variants for association with a single phenotype (gene-based association tests). Such approaches can increase statistical power by combining evidence for association over multiple phenotypes or genetic variants respectively. Canonical Correlation Analysis (CCA) measures the correlation between two sets of multidimensional variables, and thus offers the potential to combine these two approaches. To apply CCA, we must restrict the number of attributes relative to the number of samples. Hence we consider modules of genetic variation that can comprise a gene, a pathway or another biologically relevant grouping, and/or a set of phenotypes. In order to do this, we use an attribute selection strategy based on a binary genetic algorithm. Applied to a UK-based prospective cohort study of 4286 women (the British Women's Heart and Health Study), we find improved statistical power in the detection of previously reported genetic associations, and identify a number of novel pleiotropic associations between genetic variants and phenotypes. New discoveries include gene-based association of NSF with triglyceride levels and several genes (ACSM3, ERI2, IL18RAP, IL23RAP and NRG1) with left ventricular hypertrophy phenotypes. In multiple-phenotype analyses we find association of NRG1 with left ventricular hypertrophy phenotypes, fibrinogen and urea and pleiotropic relationships of F7 and F10 with Factor VII, Factor IX and cholesterol levels.

Nonsense mutation in coiled-coil domain containing 151 gene (CCDC151) causes primary ciliary dyskinesia.

Primary ciliary dyskinesia (PCD) is an autosomal-recessive disorder characterized by impaired ciliary function that leads to subsequent clinical phenotypes such as chronic sinopulmonary disease. PCD is also a genetically heterogeneous disorder with many single gene mutations leading to similar clinical phenotypes. Here, we present a novel PCD causal gene, coiled-coil domain containing 151 (CCDC151), which has been shown to be essential in motile cilia of many animals and other vertebrates but its effects in humans was not observed until currently. We observed a novel nonsense mutation in a homozygous state in the CCDC151 gene (NM_145045.4:c.925G>T:p.[E309*]) in a clinically diagnosed PCD patient from a consanguineous family of Arabic ancestry. The variant was absent in 238 randomly selected individuals indicating that the variant is rare and likely not to be a founder mutation. Our finding also shows that given prior knowledge from model organisms, even a single whole-exome sequence can be sufficient to discover a novel causal gene.

Associations between APOE and low-density lipoprotein cholesterol genotypes and cognitive and physical capability: the HALCyon programme.

The APOE ε2/3/4 genotype has been associated with low-density lipoprotein cholesterol (LDL-C) and Alzheimer disease. However, evidence for associations with measures of cognitive performance in adults without dementia has been mixed, as it is for physical performance. Associations may also be evident in other genotypes implicated in LDL-C levels. As part of the Healthy Ageing across the Life Course (HALCyon) collaborative research programme, genotypic information was obtained for APOE ε2/3/4, rs515135 (APOB), rs2228671 (LDLR) and rs629301 (SORT1) from eight cohorts of adults aged between 44 and 90 + years. We investigated associations with four measures of cognitive (word recall, phonemic fluency, semantic fluency and search speed) and physical capability (grip strength, get up and go/walk speed, timed chair rises and ability to balance) using meta-analyses. Overall, little evidence for associations between any of the genotypes and measures of cognitive capability was observed (e.g. pooled beta for APOE ε4 effect on semantic fluency z score = -0.02; 95 % CI = -0.05 to 0.02; p value = 0.3; n = 18,796). However, there was borderline evidence within studies that negative effects of APOE ε4 on nonverbal ability measures become more apparent with age. Few genotypic associations were observed with physical capability measures. The findings from our large investigation of middle-aged to older adults in the general population suggest that effects of APOE on cognitive capability are at most modest and are domain- and age-specific, while APOE has little influence on physical capability. In addition, other LDL-C-related genotypes have little impact on these traits.

Ranking non-synonymous single nucleotide polymorphisms based on disease concepts.

As the number of non-synonymous single nucleotide polymorphisms (nsSNPs) identified through whole-exome/whole-genome sequencing programs increases, researchers and clinicians are becoming increasingly reliant upon computational prediction algorithms designed to prioritize potential functional variants for further study. A large proportion of existing prediction algorithms are 'disease agnostic' but are nevertheless quite capable of predicting when a mutation is likely to be deleterious. However, most clinical and research applications of these algorithms relate to specific diseases and would therefore benefit from an approach that discriminates between functional variants specifically related to that disease from those which are not. In a whole-exome/whole-genome sequencing context, such an approach could substantially reduce the number of false positive candidate mutations. Here, we test this postulate by incorporating a disease-specific weighting scheme into the Functional Analysis through Hidden Markov Models (FATHMM) algorithm. When compared to traditional prediction algorithms, we observed an overall reduction in the number of false positives identified using a disease-specific approach to functional prediction across 17 distinct disease concepts/categories. Our results illustrate the potential benefits of making disease-specific predictions when prioritizing candidate variants in relation to specific diseases. A web-based implementation of our algorithm is available at http://fathmm.biocompute.org.uk.

A pathway-based data integration framework for prediction of disease progression.

MOTIVATION: Within medical research there is an increasing trend toward deriving multiple types of data from the same individual. The most effective prognostic prediction methods should use all available data, as this maximizes the amount of information used. In this article, we consider a variety of learning strategies to boost prediction performance based on the use of all available data. IMPLEMENTATION: We consider data integration via the use of multiple kernel learning supervised learning methods. We propose a scheme in which feature selection by statistical score is performed separately per data type and by pathway membership. We further consider the introduction of a confidence measure for the class assignment, both to remove some ambiguously labeled datapoints from the training data and to implement a cautious classifier that only makes predictions when the associated confidence is high. RESULTS: We use the METABRIC dataset for breast cancer, with prediction of survival at 2000 days from diagnosis. Predictive accuracy is improved by using kernels that exclusively use those genes, as features, which are known members of particular pathways. We show that yet further improvements can be made by using a range of additional kernels based on clinical covariates such as Estrogen Receptor (ER) status. Using this range of measures to improve prediction performance, we show that the test accuracy on new instances is nearly 80%, though predictions are only made on 69.2% of the patient cohort. AVAILABILITY: https://github.com/jseoane/FSMKL CONTACT: J.Seoane@bristol.ac.uk SUPPLEMENTARY INFORMATION: Supplementary data are available at Bioinformatics online.