Modeling drug combination effects via latent tensor reconstruction.

MOTIVATION: Combination therapies have emerged as a powerful treatment modality to overcome drug resistance and improve treatment efficacy. However, the number of possible drug combinations increases very rapidly with the number of individual drugs in consideration, which makes the comprehensive experimental screening infeasible in practice. Machine-learning models offer time- and cost-efficient means to aid this process by prioritizing the most effective drug combinations for further pre-clinical and clinical validation. However, the complexity of the underlying interaction patterns across multiple drug doses and in different cellular contexts poses challenges to the predictive modeling of drug combination effects. RESULTS: We introduce comboLTR, highly time-efficient method for learning complex, non-linear target functions for describing the responses of therapeutic agent combinations in various doses and cancer cell-contexts. The method is based on a polynomial regression via powerful latent tensor reconstruction. It uses a combination of recommender system-style features indexing the data tensor of response values in different contexts, and chemical and multi-omics features as inputs. We demonstrate that comboLTR outperforms state-of-the-art methods in terms of predictive performance and running time, and produces highly accurate results even in the challenging and practical inference scenario where full dose-response matrices are predicted for completely new drug combinations with no available combination and monotherapy response measurements in any training cell line. AVAILABILITY AND IMPLEMENTATION: comboLTR code is available at https://github.com/aalto-ics-kepaco/ComboLTR. SUPPLEMENTARY INFORMATION: Supplementary data are available at Bioinformatics online.

Learning with multiple pairwise kernels for drug bioactivity prediction.

Motivation: Many inference problems in bioinformatics, including drug bioactivity prediction, can be formulated as pairwise learning problems, in which one is interested in making predictions for pairs of objects, e.g. drugs and their targets. Kernel-based approaches have emerged as powerful tools for solving problems of that kind, and especially multiple kernel learning (MKL) offers promising benefits as it enables integrating various types of complex biomedical information sources in the form of kernels, along with learning their importance for the prediction task. However, the immense size of pairwise kernel spaces remains a major bottleneck, making the existing MKL algorithms computationally infeasible even for small number of input pairs. Results: We introduce pairwiseMKL, the first method for time- and memory-efficient learning with multiple pairwise kernels. pairwiseMKL first determines the mixture weights of the input pairwise kernels, and then learns the pairwise prediction function. Both steps are performed efficiently without explicit computation of the massive pairwise matrices, therefore making the method applicable to solving large pairwise learning problems. We demonstrate the performance of pairwiseMKL in two related tasks of quantitative drug bioactivity prediction using up to 167 995 bioactivity measurements and 3120 pairwise kernels: (i) prediction of anticancer efficacy of drug compounds across a large panel of cancer cell lines; and (ii) prediction of target profiles of anticancer compounds across their kinome-wide target spaces. We show that pairwiseMKL provides accurate predictions using sparse solutions in terms of selected kernels, and therefore it automatically identifies also data sources relevant for the prediction problem. Availability and implementation: Code is available at https://github.com/aalto-ics-kepaco. Supplementary information: Supplementary data are available at Bioinformatics online.

Computational-experimental approach to drug-target interaction mapping: A case study on kinase inhibitors.

Due to relatively high costs and labor required for experimental profiling of the full target space of chemical compounds, various machine learning models have been proposed as cost-effective means to advance this process in terms of predicting the most potent compound-target interactions for subsequent verification. However, most of the model predictions lack direct experimental validation in the laboratory, making their practical benefits for drug discovery or repurposing applications largely unknown. Here, we therefore introduce and carefully test a systematic computational-experimental framework for the prediction and pre-clinical verification of drug-target interactions using a well-established kernel-based regression algorithm as the prediction model. To evaluate its performance, we first predicted unmeasured binding affinities in a large-scale kinase inhibitor profiling study, and then experimentally tested 100 compound-kinase pairs. The relatively high correlation of 0.77 (p < 0.0001) between the predicted and measured bioactivities supports the potential of the model for filling the experimental gaps in existing compound-target interaction maps. Further, we subjected the model to a more challenging task of predicting target interactions for such a new candidate drug compound that lacks prior binding profile information. As a specific case study, we used tivozanib, an investigational VEGF receptor inhibitor with currently unknown off-target profile. Among 7 kinases with high predicted affinity, we experimentally validated 4 new off-targets of tivozanib, namely the Src-family kinases FRK and FYN A, the non-receptor tyrosine kinase ABL1, and the serine/threonine kinase SLK. Our sub-sequent experimental validation protocol effectively avoids any possible information leakage between the training and validation data, and therefore enables rigorous model validation for practical applications. These results demonstrate that the kernel-based modeling approach offers practical benefits for probing novel insights into the mode of action of investigational compounds, and for the identification of new target selectivities for drug repurposing applications.

metaCCA: summary statistics-based multivariate meta-analysis of genome-wide association studies using canonical correlation analysis.

MOTIVATION: A dominant approach to genetic association studies is to perform univariate tests between genotype-phenotype pairs. However, analyzing related traits together increases statistical power, and certain complex associations become detectable only when several variants are tested jointly. Currently, modest sample sizes of individual cohorts, and restricted availability of individual-level genotype-phenotype data across the cohorts limit conducting multivariate tests. RESULTS: We introduce metaCCA, a computational framework for summary statistics-based analysis of a single or multiple studies that allows multivariate representation of both genotype and phenotype. It extends the statistical technique of canonical correlation analysis to the setting where original individual-level records are not available, and employs a covariance shrinkage algorithm to achieve robustness.Multivariate meta-analysis of two Finnish studies of nuclear magnetic resonance metabolomics by metaCCA, using standard univariate output from the program SNPTEST, shows an excellent agreement with the pooled individual-level analysis of original data. Motivated by strong multivariate signals in the lipid genes tested, we envision that multivariate association testing using metaCCA has a great potential to provide novel insights from already published summary statistics from high-throughput phenotyping technologies. AVAILABILITY AND IMPLEMENTATION: Code is available at https://github.com/aalto-ics-kepaco CONTACTS: anna.cichonska@helsinki.fi or matti.pirinen@helsinki.fi SUPPLEMENTARY INFORMATION: Supplementary data are available at Bioinformatics online.

Profiling persistent tubercule bacilli from patient sputa during therapy predicts early drug efficacy.

BACKGROUND: New treatment options are needed to maintain and improve therapy for tuberculosis, which caused the death of 1.5 million people in 2013 despite potential for an 86 % treatment success rate. A greater understanding of Mycobacterium tuberculosis (M.tb) bacilli that persist through drug therapy will aid drug development programs. Predictive biomarkers for treatment efficacy are also a research priority. METHODS AND RESULTS: Genome-wide transcriptional profiling was used to map the mRNA signatures of M.tb from the sputa of 15 patients before and 3, 7 and 14 days after the start of standard regimen drug treatment. The mRNA profiles of bacilli through the first 2 weeks of therapy reflected drug activity at 3 days with transcriptional signatures at days 7 and 14 consistent with reduced M.tb metabolic activity similar to the profile of pre-chemotherapy bacilli. These results suggest that a pre-existing drug-tolerant M.tb population dominates sputum before and after early drug treatment, and that the mRNA signature at day 3 marks the killing of a drug-sensitive sub-population of bacilli. Modelling patient indices of disease severity with bacterial gene expression patterns demonstrated that both microbiological and clinical parameters were reflected in the divergent M.tb responses and provided evidence that factors such as bacterial load and disease pathology influence the host-pathogen interplay and the phenotypic state of bacilli. Transcriptional signatures were also defined that predicted measures of early treatment success (rate of decline in bacterial load over 3 days, TB test positivity at 2 months, and bacterial load at 2 months). CONCLUSIONS: This study defines the transcriptional signature of M.tb bacilli that have been expectorated in sputum after two weeks of drug therapy, characterizing the phenotypic state of bacilli that persist through treatment. We demonstrate that variability in clinical manifestations of disease are detectable in bacterial sputa signatures, and that the changing M.tb mRNA profiles 0-2 weeks into chemotherapy predict the efficacy of treatment 6 weeks later. These observations advocate assaying dynamic bacterial phenotypes through drug therapy as biomarkers for treatment success.

AI for targeted polypharmacology: The next frontier in drug discovery.

In drug discovery, targeted polypharmacology, i.e., targeting multiple molecular targets with a single drug, is redefining therapeutic design to address complex diseases. Pre-selected pharmacological profiles, as exemplified in kinase drugs, promise enhanced efficacy and reduced toxicity. Historically, many of such drugs were discovered serendipitously, limiting predictability and efficacy, but currently artificial intelligence (AI) offers a transformative solution. Machine learning and deep learning techniques enable modeling protein structures, generating novel compounds, and decoding their polypharmacological effects, opening an avenue for more systematic and predictive multi-target drug design. This review explores the use of AI in identifying synergistic co-targets and delineating them from anti-targets that lead to adverse effects, and then discusses advances in AI-enabled docking, generative chemistry, and proteochemometric modeling of proteome-wide compound interactions, in the context of polypharmacology. We also provide insights into challenges ahead.

Investigating the conformational landscape of AlphaFold2-predicted protein kinase structures.

Summary: Protein kinases are a family of signaling proteins, crucial for maintaining cellular homeostasis. When dysregulated, kinases drive the pathogenesis of several diseases, and are thus one of the largest target categories for drug discovery. Kinase activity is tightly controlled by switching through several active and inactive conformations in their catalytic domain. Kinase inhibitors have been designed to engage kinases in specific conformational states, where each conformation presents a unique physico-chemical environment for therapeutic intervention. Thus, modeling kinases across conformations can enable the design of novel and optimally selective kinase drugs. Due to the recent success of AlphaFold2 in accurately predicting the 3D structure of proteins based on sequence, we investigated the conformational landscape of protein kinases as modeled by AlphaFold2. We observed that AlphaFold2 is able to model several kinase conformations across the kinome, however, certain conformations are only observed in specific kinase families. Furthermore, we show that the per residue predicted local distance difference test can capture information describing structural flexibility of kinases. Finally, we evaluated the docking performance of AlphaFold2 kinase structures for enriching known ligands. Taken together, we see an opportunity to leverage AlphaFold2 models for structure-based drug discovery against kinases across several pharmacologically relevant conformational states. Availability and implementation: All code used in the analysis is freely available at https://github.com/Harmonic-Discovery/AF2-kinase-conformational-landscape.

Atlas of plasma NMR biomarkers for health and disease in 118,461 individuals from the UK Biobank.

Blood lipids and metabolites are markers of current health and future disease risk. Here, we describe plasma nuclear magnetic resonance (NMR) biomarker data for 118,461 participants in the UK Biobank. The biomarkers cover 249 measures of lipoprotein lipids, fatty acids, and small molecules such as amino acids, ketones, and glycolysis metabolites. We provide an atlas of associations of these biomarkers to prevalence, incidence, and mortality of over 700 common diseases ( nightingalehealth.com/atlas ). The results reveal a plethora of biomarker associations, including susceptibility to infectious diseases and risk of various cancers, joint disorders, and mental health outcomes, indicating that abundant circulating lipids and metabolites are risk markers beyond cardiometabolic diseases. Clustering analyses indicate similar biomarker association patterns across different disease types, suggesting latent systemic connectivity in the susceptibility to a diverse set of diseases. This work highlights the value of NMR based metabolic biomarker profiling in large biobanks for public health research and translation.

Integrative analysis of metabolite GWAS illuminates the molecular basis of pleiotropy and genetic correlation.

Pleiotropy and genetic correlation are widespread features in genome-wide association studies (GWAS), but they are often difficult to interpret at the molecular level. Here, we perform GWAS of 16 metabolites clustered at the intersection of amino acid catabolism, glycolysis, and ketone body metabolism in a subset of UK Biobank. We utilize the well-documented biochemistry jointly impacting these metabolites to analyze pleiotropic effects in the context of their pathways. Among the 213 lead GWAS hits, we find a strong enrichment for genes encoding pathway-relevant enzymes and transporters. We demonstrate that the effect directions of variants acting on biology between metabolite pairs often contrast with those of upstream or downstream variants as well as the polygenic background. Thus, we find that these outlier variants often reflect biology local to the traits. Finally, we explore the implications for interpreting disease GWAS, underscoring the potential of unifying biochemistry with dense metabolomics data to understand the molecular basis of pleiotropy in complex traits and diseases.

Effects of general and central adiposity on circulating lipoprotein, lipid, and metabolite levels in UK Biobank: A multivariable Mendelian randomization study.

Background: The direct effects of general adiposity (body mass index (BMI)) and central adiposity (waist-to-hip-ratio (WHR)) on circulating lipoproteins, lipids, and metabolites are unknown. Methods: We used new metabolic data from UK Biobank (N=109,532, a five-fold higher N over previous studies). EDTA-plasma was used to quantify 249 traits with nuclear-magnetic-resonance spectroscopy including subclass-specific lipoprotein concentrations and lipid content, plus pre-glycemic and inflammatory metabolites. We used univariable and multivariable two-stage least-squares regression models with genetic risk scores for BMI and WHR as instruments to estimate total (unadjusted) and direct (mutually-adjusted) effects of BMI and WHR on metabolic traits; plus effects on statin use and interaction by sex, statin use, and age (proxy for medication use). Findings: Higher BMI decreased apolipoprotein B and low-density lipoprotein cholesterol (LDL-C) before and after WHR-adjustment, whilst BMI increased triglycerides only before WHR-adjustment. These effects of WHR were larger and BMI-independent. Direct effects differed markedly by sex, e.g., triglycerides increased only with BMI among men, and only with WHR among women. Adiposity measures increased statin use and showed metabolic effects which differed by statin use and age. Among the youngest (38-53y, statins-5%), BMI and WHR (per-SD) increased LDL-C (total effects: 0.04-SD, 95%CI=-0.01,0.08 and 0.10-SD, 95%CI=0.02,0.17 respectively), but only WHR directly. Among the oldest (63-73y, statins-29%), BMI and WHR directly lowered LDL-C (-0.19-SD, 95%CI=-0.27,-0.11 and -0.05-SD, 95%CI=-0.16,0.06 respectively). Interpretation: Excess adiposity likely raises atherogenic lipid and metabolite levels exclusively via adiposity stored centrally, particularly among women. Apparent effects of adiposity on lowering LDL-C are likely explained by an effect of adiposity on statin use. Funding: UK Medical Research Council; British Heart Foundation; Novo Nordisk; National Institute for Health Research; Wellcome Trust; Cancer Research UK.

Circulating Metabolic Biomarkers Are Consistently Associated With Type 2 Diabetes Risk in Asian and European Populations.

CONTEXT: While Asians have a higher risk of type 2 diabetes (T2D) than Europeans for a given body mass index (BMI), it remains unclear whether the same markers of metabolic pathways are associated with diabetes. OBJECTIVE: We evaluated associations between metabolic biomarkers and incidence of T2D in 3 major Asian ethnic groups (Chinese, Malay, and Indian) and a European population. METHODS: We analyzed data from adult males and females of 2 cohorts from Singapore (n = 6393) consisting of Chinese, Malays, and Indians and 3 cohorts of European-origin participants from Finland (n = 14 558). We used nuclear magnetic resonance to quantify 154 circulating metabolic biomarkers at baseline and performed logistic regression to assess associations with T2D risk adjusted for age, sex, BMI and glycemic markers. RESULTS: Of the 154 metabolic biomarkers, 59 were associated with higher risk of T2D in both Asians and Europeans (P < 0.0003, Bonferroni-corrected). These included branched chain and aromatic amino acids, the inflammatory marker glycoprotein acetyls, total fatty acids, monounsaturated fatty acids, apolipoprotein B, larger very low-density lipoprotein particle sizes, and triglycerides. In addition, 13 metabolites were associated with a lower T2D risk in both populations, including omega-6 polyunsaturated fatty acids and larger high-density lipoprotein particle sizes. Associations were consistent within the Asian ethnic groups (all Phet ≥ 0.05) and largely consistent for the Asian and European populations (Phet ≥ 0.05 for 128 of 154 metabolic biomarkers). CONCLUSION: Metabolic biomarkers across several biological pathways were consistently associated with T2D risk in Asians and Europeans.

Metabolic biomarker profiling for identification of susceptibility to severe pneumonia and COVID-19 in the general population.

Biomarkers of low-grade inflammation have been associated with susceptibility to a severe infectious disease course, even when measured prior to disease onset. We investigated whether metabolic biomarkers measured by nuclear magnetic resonance (NMR) spectroscopy could be associated with susceptibility to severe pneumonia (2507 hospitalised or fatal cases) and severe COVID-19 (652 hospitalised cases) in 105,146 generally healthy individuals from UK Biobank, with blood samples collected 2007-2010. The overall signature of metabolic biomarker associations was similar for the risk of severe pneumonia and severe COVID-19. A multi-biomarker score, comprised of 25 proteins, fatty acids, amino acids, and lipids, was associated equally strongly with enhanced susceptibility to severe COVID-19 (odds ratio 2.9 [95%CI 2.1-3.8] for highest vs lowest quintile) and severe pneumonia events occurring 7-11 years after blood sampling (2.6 [1.7-3.9]). However, the risk for severe pneumonia occurring during the first 2 years after blood sampling for people with elevated levels of the multi-biomarker score was over four times higher than for long-term risk (8.0 [4.1-15.6]). If these hypothesis generating findings on increased susceptibility to severe pneumonia during the first few years after blood sampling extend to severe COVID-19, metabolic biomarker profiling could potentially complement existing tools for identifying individuals at high risk. These results provide novel molecular understanding on how metabolic biomarkers reflect the susceptibility to severe COVID-19 and other infections in the general population.

National policies for mitigating the COVID-19 pandemic include stricter measures for people considered to be at high risk of severe and potentially fatal cases of the disease. Although older age and pre-existing health conditions are strong risk factors, it is poorly understood why susceptibility varies so widely in the population. People with cardiometabolic diseases, such as diabetes and liver diseases, or chronic inflammation are at higher risk of severe COVID-19 and other infections including pneumonia. These conditions alter the molecules circulating in the blood, providing potential 'biomarkers' to determine whether a person is more likely to develop a fatal infection. Uncovering these blood biomarkers could help to identify people who are prone to life-threatening infections despite not having ever been diagnosed with a cardiometabolic disease. To find these biomarkers, Julkunen et al. studied blood samples that had been collected from 105,000 healthy individuals in the United Kingdom over ten years ago. The data showed that individuals with biomarkers linked to low-grade inflammation and cardiometabolic disease were more likely to have died or been hospitalised with pneumonia. A score based on 25 of these biomarkers provided the best predictor of severe pneumonia. This biomarker score performed up to four times better within the first few years after blood sampling compared to predicting cases of pneumonia a decade later. The same blood biomarker changes were also linked with developing severe COVID-19 over ten years after the blood samples had been collected. The predictive value of the biomarker score was similar for both severe COVID-19 and the long-term risk of severe pneumonia. Julkunen et al. propose that the metabolic biomarkers reflect inhibited immunity that impairs response to infections. The results from over 100,000 individuals suggest that these blood biomarkers may help to identify people at high risk of severe COVID-19 or other infectious diseases.

An expanded analysis framework for multivariate GWAS connects inflammatory biomarkers to functional variants and disease.

Multivariate methods are known to increase the statistical power to detect associations in the case of shared genetic basis between phenotypes. They have, however, lacked essential analytic tools to follow-up and understand the biology underlying these associations. We developed a novel computational workflow for multivariate GWAS follow-up analyses, including fine-mapping and identification of the subset of traits driving associations (driver traits). Many follow-up tools require univariate regression coefficients which are lacking from multivariate results. Our method overcomes this problem by using Canonical Correlation Analysis to turn each multivariate association into its optimal univariate Linear Combination Phenotype (LCP). This enables an LCP-GWAS, which in turn generates the statistics required for follow-up analyses. We implemented our method on 12 highly correlated inflammatory biomarkers in a Finnish population-based study. Altogether, we identified 11 associations, four of which (F5, ABO, C1orf140 and PDGFRB) were not detected by biomarker-specific analyses. Fine-mapping identified 19 signals within the 11 loci and driver trait analysis determined the traits contributing to the associations. A phenome-wide association study on the 19 representative variants from the signals in 176,899 individuals from the FinnGen study revealed 53 disease associations (p < 1 × 10-4). Several reported pQTLs in the 11 loci provided orthogonal evidence for the biologically relevant functions of the representative variants. Our novel multivariate analysis workflow provides a powerful addition to standard univariate GWAS analyses by enabling multivariate GWAS follow-up and thus promoting the advancement of powerful multivariate methods in genomics.

Crowdsourced mapping of unexplored target space of kinase inhibitors.

Despite decades of intensive search for compounds that modulate the activity of particular protein targets, a large proportion of the human kinome remains as yet undrugged. Effective approaches are therefore required to map the massive space of unexplored compound-kinase interactions for novel and potent activities. Here, we carry out a crowdsourced benchmarking of predictive algorithms for kinase inhibitor potencies across multiple kinase families tested on unpublished bioactivity data. We find the top-performing predictions are based on various models, including kernel learning, gradient boosting and deep learning, and their ensemble leads to a predictive accuracy exceeding that of single-dose kinase activity assays. We design experiments based on the model predictions and identify unexpected activities even for under-studied kinases, thereby accelerating experimental mapping efforts. The open-source prediction algorithms together with the bioactivities between 95 compounds and 295 kinases provide a resource for benchmarking prediction algorithms and for extending the druggable kinome.

Leveraging multi-way interactions for systematic prediction of pre-clinical drug combination effects.

We present comboFM, a machine learning framework for predicting the responses of drug combinations in pre-clinical studies, such as those based on cell lines or patient-derived cells. comboFM models the cell context-specific drug interactions through higher-order tensors, and efficiently learns latent factors of the tensor using powerful factorization machines. The approach enables comboFM to leverage information from previous experiments performed on similar drugs and cells when predicting responses of new combinations in so far untested cells; thereby, it achieves highly accurate predictions despite sparsely populated data tensors. We demonstrate high predictive performance of comboFM in various prediction scenarios using data from cancer cell line pharmacogenomic screens. Subsequent experimental validation of a set of previously untested drug combinations further supports the practical and robust applicability of comboFM. For instance, we confirm a novel synergy between anaplastic lymphoma kinase (ALK) inhibitor crizotinib and proteasome inhibitor bortezomib in lymphoma cells. Overall, our results demonstrate that comboFM provides an effective means for systematic pre-screening of drug combinations to support precision oncology applications.

Identification of drug candidates and repurposing opportunities through compound-target interaction networks.

INTRODUCTION: System-wide identification of both on- and off-targets of chemical probes provides improved understanding of their therapeutic potential and possible adverse effects, thereby accelerating and de-risking drug discovery process. Given the high costs of experimental profiling of the complete target space of drug-like compounds, computational models offer systematic means for guiding these mapping efforts. These models suggest the most potent interactions for further experimental or pre-clinical evaluation both in cell line models and in patient-derived material. AREAS COVERED: The authors focus here on network-based machine learning models and their use in the prediction of novel compound-target interactions both in target-based and phenotype-based drug discovery applications. While currently being used mainly in complementing the experimentally mapped compound-target networks for drug repurposing applications, such as extending the target space of already approved drugs, these network pharmacology approaches may also suggest completely unexpected and novel investigational probes for drug development. EXPERT OPINION: Although the studies reviewed here have already demonstrated that network-centric modeling approaches have the potential to identify candidate compounds and selective targets in disease networks, many challenges still remain. In particular, these challenges include how to incorporate the cellular context and genetic background into the disease networks to enable more stratified and selective target predictions, as well as how to make the prediction models more realistic for the practical drug discovery and therapeutic applications.