ECOLE: Learning to call copy number variants on whole exome sequencing data

Copy number variants (CNV) are shown to contribute to the etiology of several genetic disorders. Accurate detection of CNVs on whole exome sequencing (WES) data has been a long sought-after goal for use in clinics. This was not possible despite recent improvements in performance because algorithms mostly suffer from low precision and even lower recall on expert-curated gold standard call sets. Here, we present a deep learning-based somatic and germline CNV caller for WES data, named ECOLE. Based on a variant of the transformer architecture, the model learns to call CNVs per exon, using high-confidence calls made on matched WGS samples. We further train and fine-tune the model with a small set of expert calls via transfer learning. We show that ECOLE achieves high performance on human expert labelled data for the first time with 68.7% precision and 49.6% recall. This corresponds to precision and recall improvements of 18.7% and 30.8% over the next best-performing methods, respectively. We also show that the same fine-tuning strategy using tumor samples enables ECOLE to detect RT-qPCR-validated variations in bladder cancer samples without the need for a control sample. ECOLE is available at https://github.com/ciceklab/ECOLE .

PiDeeL: metabolic pathway-informed deep learning model for survival analysis and pathological classification of gliomas

MOTIVATION

Online assessment of tumor characteristics during surgery is important and has the potential to establish an intra-operative surgeon feedback mechanism. With the availability of such feedback, surgeons could decide to be more liberal or conservative regarding the resection of the tumor. While there are methods to perform metabolomics-based tumor pathology prediction, their model complexity predictive performance is limited by the small dataset sizes. Furthermore, the information conveyed by the feedback provided on the tumor tissue could be improved both in terms of content and accuracy.

RESULTS

In this study, we propose a metabolic pathway-informed deep learning model (PiDeeL) to perform survival analysis and pathology assessment based on metabolite concentrations. We show that incorporating pathway information into the model architecture substantially reduces parameter complexity and achieves better survival analysis and pathological classification performance. With these design decisions, we show that PiDeeL improves tumor pathology prediction performance of the state-of-the-art in terms of the Area Under the ROC Curve by 3.38% and the Area Under the Precision-Recall Curve by 4.06%. Similarly, with respect to the time-dependent concordance index (c-index), PiDeeL achieves better survival analysis performance (improvement of 4.3%) when compared to the state-of-the-art. Moreover, we show that importance analyses performed on input metabolite features as well as pathway-specific neurons of PiDeeL provide insights into tumor metabolism. We foresee that the use of this model in the surgery room will help surgeons adjust the surgery plan on the fly and will result in better prognosis estimates tailored to surgical procedures.

AVAILABILITY AND IMPLEMENTATION

The code is released at https://github.com/ciceklab/PiDeeL. The data used in this study are released at https://zenodo.org/record/7228791.

DeepSide: A Deep Learning Approach for Drug Side Effect Prediction

Drug failures due to unforeseen adverse effects at clinical trials pose health risks for the participants and lead to substantial financial losses. Side effect prediction algorithms have the potential to guide the drug design process. LINCS L1000 dataset provides a vast resource of cell line gene expression data perturbed by different drugs and creates a knowledge base for context specific features. The state-of-the-art approach that aims at using context specific information relies on only the high-quality experiments in LINCS L1000 and discards a large portion of the experiments. In this study, our goal is to boost the prediction performance by utilizing this data to its full extent. We experiment with 5 deep learning architectures. We find that a multi-modal architecture produces the best predictive performance among multi-layer perceptron-based architectures when drug chemical structure (CS), and the full set of drug perturbed gene expression profiles (GEX) are used as modalities. Overall, we observe that the CS is more informative than the GEX. A convolutional neural network-based model that uses only SMILES string representation of the drugs achieves the best results and provides 13.0% macro-AUC and 3.1% micro-AUC improvements over the state-of-the-art. We also show that the model is able to predict side effect-drug pairs that are reported in the literature but was missing in the ground truth side effect dataset. DeepSide is available at http://github.com/OnurUner/DeepSide.

A unifying network modeling approach for codon optimization

MOTIVATION

Synthesizing genes to be expressed in other organisms is an essential tool in biotechnology. While the many-to-one mapping from codons to amino acids makes the genetic code degenerate, codon usage in a particular organism is not random either. This bias in codon use may have a remarkable effect on the level of gene expression. A number of measures have been developed to quantify a given codon sequence's strength to express a gene in a host organism. Codon optimization aims to find a codon sequence that will optimize one or more of these measures. Efficient computational approaches are needed since the possible number of codon sequences grows exponentially as the number of amino acids increases.

RESULTS

We develop a unifying modeling approach for codon optimization. With our mathematical formulations based on graph/network representations of amino acid sequences, any combination of measures can be optimized in the same framework by finding a path satisfying additional limitations in an acyclic layered network. We tested our approach on bi-objectives commonly used in the literature, namely, Codon Pair Bias versus Codon Adaptation Index and Relative Codon Pair Bias versus Relative Codon Bias. However, our framework is general enough to handle any number of objectives concurrently with certain restrictions or preferences on the use of specific nucleotide sequences. We implemented our models using Python's Gurobi interface and showed the efficacy of our approach even for the largest proteins available. We also provided experimentation showing that highly expressed genes have objective values close to the optimized values in the bi-objective codon design problem.

AVAILABILITY AND IMPLEMENTATION

http://alpersen.bilkent.edu.tr/NetworkCodon.zip.

SUPPLEMENTARY INFORMATION

Supplementary data are available at Bioinformatics online.

DeepND: Deep multitask learning of gene risk for comorbid neurodevelopmental disorders

Autism spectrum disorder and intellectual disability are comorbid neurodevelopmental disorders with complex genetic architectures. Despite large-scale sequencing studies, only a fraction of the risk genes was identified for both. We present a network-based gene risk prioritization algorithm, DeepND, that performs cross-disorder analysis to improve prediction by exploiting the comorbidity of autism spectrum disorder (ASD) and intellectual disability (ID) via multitask learning. Our model leverages information from human brain gene co-expression networks using graph convolutional networks, learning which spatiotemporal neurodevelopmental windows are important for disorder etiologies and improving the state-of-the-art prediction in single- and cross-disorder settings. DeepND identifies the prefrontal and motor-somatosensory cortex (PFC-MFC) brain region and periods from early- to mid-fetal and from early childhood to young adulthood as the highest neurodevelopmental risk windows for ASD and ID. We investigate ASD- and ID-associated copy-number variation (CNV) regions and report our findings for several susceptibility gene candidates. DeepND can be generalized to analyze any combinations of comorbid disorders.

•

DeepND can co-analyze comorbid neurodevelopmental disorders to discover risk genes

•

The approach employs multitask learning to learn shared and disorder-specific weights

•

DeepND uses graph convolution to process gene interactions in multiple networks

•

The model includes a mixture-of-experts model to detect informative networks

While risk-gene-discovery algorithms have complemented exome/genome-sequencing studies of neurodevelopmental disorders, they are not capable of co-analyzing multiple comorbid conditions like autism and intellectual disability. A common approach is analyzing disorders one by one and comparing the outcomes. With this approach, the method does not utilize cross-disorder interactions and is bound by limited evidence per disorder. We address this gap with a technique, Deep Neurodevelopmental Disorders (DeepND), that uses multitask learning to co-analyze data from multiple disorders to learn shared and disorder-specific patterns. DeepND includes graph convolutional neural networks that process gene-interaction information from multiple networks. DeepND also learns which networks are important for disorder etiologies. Based on this, we propose an interpretable risk-gene-discovery algorithm for neuropsychiatric disorders.

MatchMaker: A Deep Learning Framework for Drug Synergy Prediction

Drug combination therapies have been a viable strategy for the treatment of complex diseases such as cancer due to increased efficacy and reduced side effects. However, experimentally validating all possible combinations for synergistic interaction even with high-throughout screens is intractable due to vast combinatorial search space. Computational techniques can reduce the number of combinations to be evaluated experimentally by prioritizing promising candidates. We present MatchMaker that predicts drug synergy scores using drug chemical structure information and gene expression profiles of cell lines in a deep learning framework. For the first time, our model utilizes the largest known drug combination dataset to date, DrugComb. We compare the performance of MatchMaker with the state-of-the-art models and observe up to  ∼ 15% correlation and  ∼ 33% mean squared error (MSE) improvements over the next best method. We investigate the cell types and drug pairs that are relatively harder to predict and present novel candidate pairs. MatchMaker is built and available at https://github.com/tastanlab/matchmaker.

Targeted metabolomics analyses for brain tumor margin assessment during surgery

MOTIVATION

Identification and removal of micro-scale residual tumor tissue during brain tumor surgery are key for survival in glioma patients. For this goal, High-Resolution Magic Angle Spinning Nuclear Magnetic Resonance (HRMAS NMR) spectroscopy-based assessment of tumor margins during surgery has been an effective method. However, the time required for metabolite quantification and the need for human experts such as a pathologist to be present during surgery are major bottlenecks of this technique. While machine learning techniques that analyze the NMR spectrum in an untargeted manner (i.e. using the full raw signal) have been shown to effectively automate this feedback mechanism, high dimensional and noisy structure of the NMR signal limits the attained performance.

RESULTS

In this study, we show that identifying informative regions in the HRMAS NMR spectrum and using them for tumor margin assessment improves the prediction power. We use the spectra normalized with the ERETIC (electronic reference to access in vivo concentrations) method which uses an external reference signal to calibrate the HRMAS NMR spectrum. We train models to predict quantities of metabolites from annotated regions of this spectrum. Using these predictions for tumor margin assessment provides performance improvements up to 4.6% the Area Under the ROC Curve (AUC-ROC) and 2.8% the Area Under the Precision-Recall Curve (AUC-PR). We validate the importance of various tumor biomarkers and identify a novel region between 7.97 ppm and 8.09 ppm as a new candidate for a glioma biomarker.

AVAILABILITY AND IMPLEMENTATION

The code is released at https://github.com/ciceklab/targeted_brain_tumor_margin_assessment. The data underlying this article are available in Zenodo, at https://doi.org/10.5281/zenodo.5781769.

SUPPLEMENTARY INFORMATION

Supplementary data are available at Bioinformatics online.

From cell lines to cancer patients: personalized drug synergy prediction

MOTIVATION

Combination drug therapies are effective treatments for cancer. However, the genetic heterogeneity of the patients and exponentially large space of drug pairings pose significant challenges for finding the right combination for a specific patient. Current in silico prediction methods can be instrumental in reducing the vast number of candidate drug combinations. However, existing powerful methods are trained with cancer cell line gene expression data, which limits their applicability in clinical settings. While synergy measurements on cell line models are available at large scale, patient-derived samples are too few to train a complex model. On the other hand, patient-specific single-drug response data are relatively more available.

RESULTS

In this work, we propose a deep learning framework, Personalized Deep Synergy Predictor (PDSP), that enables us to use the patient-specific single drug response data for customizing patient drug synergy predictions. PDSP is first trained to learn synergy scores of drug pairs and their single drug responses for a given cell line using drug structures and large scale cell line gene expression data. Then, the model is fine-tuned for patients with their patient gene expression data and associated single drug response measured on the patient ex vivo samples. In this study, we evaluate PDSP on data from three leukemia patients and observe that it improves the prediction accuracy by 27% compared to models trained on cancer cell line data.

AVAILABILITY AND IMPLEMENTATION

PDSP is available at https://github.com/hikuru/PDSP.

AMULET: a novel read count-based method for effective multiplet detection from single nucleus ATAC-seq data

Detecting multiplets in single nucleus (sn)ATAC-seq data is challenging due to data sparsity and limited dynamic range. AMULET (ATAC-seq MULtiplet Estimation Tool) enumerates regions with greater than two uniquely aligned reads across the genome to effectively detect multiplets. We evaluate the method by generating snATAC-seq data in the human blood and pancreatic islet samples. AMULET has high precision, estimated via donor-based multiplexing, and high recall, estimated via simulated multiplets, compared to alternatives and identifies multiplets most effectively when a certain read depth of 25K median valid reads per nucleus is achieved.

DORMAN: Database of Reconstructed MetAbolic Networks

Genome-scale reconstructed metabolic networks have provided an organism specific understanding of cellular processes and their relations to phenotype. As they are deemed essential to study metabolism, the number of organisms with reconstructed metabolic networks continues to increase. This everlasting research interest lead to the development of online systems/repositories that store existing reconstructions and enable new model generation, integration, and constraint-based analyses. While features that support model reconstruction are widely available, current systems lack the means to help users who are interested in analyzing the topology of the reconstructed networks. Here, we present the Database of Reconstructed Metabolic Networks - DORMAN. DORMAN is a centralized online database that stores SBML-based reconstructed metabolic networks published in the literature, and provides web-based computational tools for visualizing and analyzing the model topology. Novel features of DORMAN are (i) interactive visualization interface that allows rendering of the complete network as well as editing and exporting the model, (ii) hierarchical navigation that provides efficient access to connected entities in the model, (iii) built-in query interface that allow posing topological queries, and finally, and (iv) model comparison tool that enables comparing models with different nomenclatures, using approximate string matching. DORMAN is online and freely accessible at http://ciceklab.cs.bilkent.edu.tr/dorman.

SPADIS: An Algorithm for Selecting Predictive and Diverse SNPs in GWAS

Phenotypic heritability of complex traits and diseases is seldom explained by individual genetic variants identified in genome-wide association studies (GWAS). Many methods have been developed to select a subset of variant loci, which are associated with or predictive of the phenotype. Selecting connected SNPs on SNP-SNP networks have been proven successful in finding biologically interpretable and predictive SNPs. However, we argue that the connectedness constraint favors selecting redundant features that affect similar biological processes and therefore does not necessarily yield better predictive performance. In this paper, we propose a novel method called SPADIS that favors the selection of remotely located SNPs in order to account for their complementary effects in explaining a phenotype. SPADIS selects a diverse set of loci on a SNP-SNP network. This is achieved by maximizing a submodular set function with a greedy algorithm that ensures a constant factor approximation to the optimal solution. We compare SPADIS to the state-of-the-art method SConES, on a dataset of Arabidopsis Thaliana with continuous flowering time phenotypes. SPADIS has better average phenotype prediction performance in 15 out of 17 phenotypes when the same number of SNPs are selected and provides consistent improvements across multiple networks and settings on average. Moreover, it identifies more candidate genes and runs faster.

Whole-Genome and RNA Sequencing Reveal Variation and Transcriptomic Coordination in the Developing Human Prefrontal Cortex

Gene expression levels vary across developmental stage, cell type, and region in the brain. Genomic variants also contribute to the variation in expression, and some neuropsychiatric disorder loci may exert their effects through this mechanism. To investigate these relationships, we present BrainVar, a unique resource of paired whole-genome and bulk tissue RNA sequencing from the dorsolateral prefrontal cortex of 176 individuals across prenatal and postnatal development. Here we identify common variants that alter gene expression (expression quantitative trait loci [eQTLs]) constantly across development or predominantly during prenatal or postnatal stages. Both "constant" and "temporal-predominant" eQTLs are enriched for loci associated with neuropsychiatric traits and disorders and colocalize with specific variants. Expression levels of more than 12,000 genes rise or fall in a concerted late-fetal transition, with the transitional genes enriched for cell-type-specific genes and neuropsychiatric risk loci, underscoring the importance of cataloging developmental trajectories in understanding cortical physiology and pathology.

Genome Reconstruction Attacks Against Genomic Data-Sharing Beacons

Sharing genome data in a privacy-preserving way stands as a major bottleneck in front of the scientific progress promised by the big data era in genomics. A community-driven protocol named genomic data-sharing beacon protocol has been widely adopted for sharing genomic data. The system aims to provide a secure, easy to implement, and standardized interface for data sharing by only allowing yes/no queries on the presence of specific alleles in the dataset. However, beacon protocol was recently shown to be vulnerable against membership inference attacks. In this paper, we show that privacy threats against genomic data sharing beacons are not limited to membership inference. We identify and analyze a novel vulnerability of genomic data-sharing beacons: genome reconstruction. We show that it is possible to successfully reconstruct a substantial part of the genome of a victim when the attacker knows the victim has been added to the beacon in a recent update. In particular, we show how an attacker can use the inherent correlations in the genome and clustering techniques to run such an attack in an efficient and accurate way. We also show that even if multiple individuals are added to the beacon during the same update, it is possible to identify the victim's genome with high confidence using traits that are easily accessible by the attacker (e.g., eye color or hair type). Moreover, we show how a reconstructed genome using a beacon that is not associated with a sensitive phenotype can be used for membership inference attacks to beacons with sensitive phenotypes (e.g., HIV+). The outcome of this work will guide beacon operators on when and how to update the content of the beacon and help them (along with the beacon participants) make informed decisions.

The effect of kinship in re-identification attacks against genomic data sharing beacons

MOTIVATION

Big data era in genomics promises a breakthrough in medicine, but sharing data in a private manner limit the pace of field. Widely accepted 'genomic data sharing beacon' protocol provides a standardized and secure interface for querying the genomic datasets. The data are only shared if the desired information (e.g. a certain variant) exists in the dataset. Various studies showed that beacons are vulnerable to re-identification (or membership inference) attacks. As beacons are generally associated with sensitive phenotype information, re-identification creates a significant risk for the participants. Unfortunately, proposed countermeasures against such attacks have failed to be effective, as they do not consider the utility of beacon protocol.

RESULTS

In this study, for the first time, we analyze the mitigation effect of the kinship relationships among beacon participants against re-identification attacks. We argue that having multiple family members in a beacon can garble the information for attacks since a substantial number of variants are shared among kin-related people. Using family genomes from HapMap and synthetically generated datasets, we show that having one of the parents of a victim in the beacon causes (i) significant decrease in the power of attacks and (ii) substantial increase in the number of queries needed to confirm an individual's beacon membership. We also show how the protection effect attenuates when more distant relatives, such as grandparents are included alongside the victim. Furthermore, we quantify the utility loss due adding relatives and show that it is smaller compared with flipping based techniques.

Apollo: a sequencing-technology-independent, scalable and accurate assembly polishing algorithm

MOTIVATION

Third-generation sequencing technologies can sequence long reads that contain as many as 2 million base pairs. These long reads are used to construct an assembly (i.e. the subject's genome), which is further used in downstream genome analysis. Unfortunately, third-generation sequencing technologies have high sequencing error rates and a large proportion of base pairs in these long reads is incorrectly identified. These errors propagate to the assembly and affect the accuracy of genome analysis. Assembly polishing algorithms minimize such error propagation by polishing or fixing errors in the assembly by using information from alignments between reads and the assembly (i.e. read-to-assembly alignment information). However, current assembly polishing algorithms can only polish an assembly using reads from either a certain sequencing technology or a small assembly. Such technology-dependency and assembly-size dependency require researchers to (i) run multiple polishing algorithms and (ii) use small chunks of a large genome to use all available readsets and polish large genomes, respectively.

RESULTS

We introduce Apollo, a universal assembly polishing algorithm that scales well to polish an assembly of any size (i.e. both large and small genomes) using reads from all sequencing technologies (i.e. second- and third-generation). Our goal is to provide a single algorithm that uses read sets from all available sequencing technologies to improve the accuracy of assembly polishing and that can polish large genomes. Apollo (i) models an assembly as a profile hidden Markov model (pHMM), (ii) uses read-to-assembly alignment to train the pHMM with the Forward-Backward algorithm and (iii) decodes the trained model with the Viterbi algorithm to produce a polished assembly. Our experiments with real readsets demonstrate that Apollo is the only algorithm that (i) uses reads from any sequencing technology within a single run and (ii) scales well to polish large assemblies without splitting the assembly into multiple parts.

AVAILABILITY AND IMPLEMENTATION

Source code is available at https://github.com/CMU-SAFARI/Apollo.

SUPPLEMENTARY INFORMATION

Supplementary data are available at Bioinformatics online.

A Tool for Detecting Complementary Single Nucleotide Polymorphism Pairs in Genome-Wide Association Studies for Epistasis Testing

Detecting interacting loci pairs has been instrumental to understand disease etiology when single locus associations do not fully account for the underlying heritability. However, the number of loci to test is prohibitively large. Epistasis test prioritization algorithms rank likely epistatic single nucleotide polymorphism (SNP) pairs to limit the number of statistical tests. Potpourri detects epistatic SNP pairs by diversifying the selected SNPs' genomic regions and investigating their co-occurrence patterns over the case cohort. It can also input and further prioritize SNPs in regulatory or coding regions. The program identifies and returns a list of prioritized SNP pairs for epistasis testing. This article describes how to use the program and the details of the input and output data.

Potpourri: An Epistasis Test Prioritization Algorithm via Diverse SNP Selection

Genome-wide association studies (GWAS) explain a fraction of the underlying heritability of genetic diseases. Investigating epistatic interactions between two or more loci help to close this gap. Unfortunately, the sheer number of loci combinations to process and hypotheses prohibit the process both computationally and statistically. Epistasis test prioritization algorithms rank likely epistatic single nucleotide polymorphism (SNP) pairs to limit the number of tests. However, they still suffer from very low precision. It was shown in the literature that selecting SNPs that are individually correlated with the phenotype and also diverse with respect to genomic location leads to better phenotype prediction due to genetic complementation. Here, we propose that an algorithm that pairs SNPs from such diverse regions and ranks them can improve prediction power. We propose an epistasis test prioritization algorithm that optimizes a submodular set function to select a diverse and complementary set of genomic regions that span the underlying genome. The SNP pairs from these regions are then further ranked w.r.t. their co-coverage of the case cohort. We compare our algorithm with the state of the art on three GWAS and show that (1) we substantially improve precision (from 0.003 to 0.652) while maintaining the significance of selected pairs, (2) decrease the number of tests by 25-fold, and (3) decrease the runtime by 4-fold. We also show that promoting SNPs from regulatory/coding regions improves the performance (up to 0.8). Potpourri is available at http:/ciceklab.cs.bilkent.edu.tr/potpourri.

Machine learning assisted intraoperative assessment of brain tumor margins using HRMAS NMR spectroscopy

Complete resection of the tumor is important for survival in glioma patients. Even if the gross total resection was achieved, left-over micro-scale tissue in the excision cavity risks recurrence. High Resolution Magic Angle Spinning Nuclear Magnetic Resonance (HRMAS NMR) technique can distinguish healthy and malign tissue efficiently using peak intensities of biomarker metabolites. The method is fast, sensitive and can work with small and unprocessed samples, which makes it a good fit for real-time analysis during surgery. However, only a targeted analysis for the existence of known tumor biomarkers can be made and this requires a technician with chemistry background, and a pathologist with knowledge on tumor metabolism to be present during surgery. Here, we show that we can accurately perform this analysis in real-time and can analyze the full spectrum in an untargeted fashion using machine learning. We work on a new and large HRMAS NMR dataset of glioma and control samples (n = 565), which are also labeled with a quantitative pathology analysis. Our results show that a random forest based approach can distinguish samples with tumor cells and controls accurately and effectively with a median AUC of 85.6% and AUPR of 93.4%. We also show that we can further distinguish benign and malignant samples with a median AUC of 87.1% and AUPR of 96.1%. We analyze the feature (peak) importance for classification to interpret the results of the classifier. We validate that known malignancy biomarkers such as creatine and 2-hydroxyglutarate play an important role in distinguishing tumor and normal cells and suggest new biomarker regions. The code is released at http://github.com/ciceklab/HRMAS_NC.

ST-Steiner: a spatio-temporal gene discovery algorithm

MOTIVATION

Whole exome sequencing (WES) studies for autism spectrum disorder (ASD) could identify only around six dozen risk genes to date because the genetic architecture of the disorder is highly complex. To speed the gene discovery process up, a few network-based ASD gene discovery algorithms were proposed. Although these methods use static gene interaction networks, functional clustering of genes is bound to evolve during neurodevelopment and disruptions are likely to have a cascading effect on the future associations. Thus, approaches that disregard the dynamic nature of neurodevelopment are limited.

RESULTS

Here, we present a spatio-temporal gene discovery algorithm, which leverages information from evolving gene co-expression networks of neurodevelopment. The algorithm solves a prize-collecting Steiner forest-based problem on co-expression networks, adapted to model neurodevelopment and transfer information from precursor neurodevelopmental windows. The decisions made by the algorithm can be traced back, adding interpretability to the results. We apply the algorithm on ASD WES data of 3871 samples and identify risk clusters using BrainSpan co-expression networks of early- and mid-fetal periods. On an independent dataset, we show that incorporation of the temporal dimension increases the predictive power: predicted clusters are hit more and show higher enrichment in ASD-related functions compared with the state-of-the-art.

AVAILABILITY AND IMPLEMENTATION

The code is available at http://ciceklab.cs.bilkent.edu.tr/st-steiner.

SUPPLEMENTARY INFORMATION

Supplementary data are available at Bioinformatics online.

Predicting Carbon Spectrum in Heteronuclear Single Quantum Coherence Spectroscopy for Online Feedback During Surgery

¹H High-Resolution Magic Angle Spinning (HRMAS) Nuclear Magnetic Resonance (NMR) is a reliable technology used for detecting metabolites in solid tissues. Fast response time enables guiding surgeons in real time, for detecting tumor cells that are left over in the excision cavity. However, severe overlap of spectral resonances in 1D signal often render distinguishing metabolites impossible. In that case, Heteronuclear Single Quantum Coherence Spectroscopy (HSQC) NMR is applied which can distinguish metabolites by generating 2D spectra ( ¹H- ¹³C). Unfortunately, this analysis requires much longer time and prohibits real time analysis. Thus, obtaining 2D spectrum fast has major implications in medicine. In this study, we show that using multiple multivariate regression and statistical total correlation spectroscopy, we can learn the relation between the ¹H and ¹³C dimensions. Learning is possible with small sample sizes and without the need for performing the HSQC analysis, we can predict the ¹³C dimension by just performing ¹H HRMAS NMR experiment. We show on a rat model of central nervous system tissues (80 samples, 5 tissues) that our methods achieve 0.971 and 0.957 mean R² values, respectively. Our tests on 15 human brain tumor samples show that we can predict 104 groups of 39 metabolites with 97 percent accuracy. Finally, we show that we can predict the presence of a drug resistant tumor biomarker (creatine) despite obstructed signal in ¹H dimension. In practice, this information can provide valuable feedback to the surgeon to further resect the cavity to avoid potential recurrence.

Large-Scale Exome Sequencing Study Implicates Both Developmental and Functional Changes in the Neurobiology of Autism

We present the largest exome sequencing study of autism spectrum disorder (ASD) to date (n = 35,584 total samples, 11,986 with ASD). Using an enhanced analytical framework to integrate de novo and case-control rare variation, we identify 102 risk genes at a false discovery rate of 0.1 or less. Of these genes, 49 show higher frequencies of disruptive de novo variants in individuals ascertained to have severe neurodevelopmental delay, whereas 53 show higher frequencies in individuals ascertained to have ASD; comparing ASD cases with mutations in these groups reveals phenotypic differences. Expressed early in brain development, most risk genes have roles in regulation of gene expression or neuronal communication (i.e., mutations effect neurodevelopmental and neurophysiological changes), and 13 fall within loci recurrently hit by copy number variants. In cells from the human cortex, expression of risk genes is enriched in excitatory and inhibitory neuronal lineages, consistent with multiple paths to an excitatory-inhibitory imbalance underlying ASD.

Metabolomics of Small Intestine Neuroendocrine Tumors and Related Hepatic Metastases

To assess the metabolomic fingerprint of small intestine neuroendocrine tumors (SI-NETs) and related hepatic metastases, and to investigate the influence of the hepatic environment on SI-NETs metabolome. Ninety-four tissue samples, including 46 SI-NETs, 18 hepatic NET metastases and 30 normal SI and liver samples, were analyzed using 1H-magic angle spinning (HRMAS) NMR nuclear magnetic resonance (NMR) spectroscopy. Twenty-seven metabolites were identified and quantified. Differences between primary NETs vs. normal SI and primary NETs vs. hepatic metastases, were assessed. Network analysis was performed according to several clinical and pathological features. Succinate, glutathion, taurine, myoinositol and glycerophosphocholine characterized NETs. Normal SI specimens showed higher levels of alanine, creatine, ethanolamine and aspartate. PLS-DA revealed a continuum-like distribution among normal SI, G1-SI-NETs and G2-SI-NETs. The G2-SI-NET distribution was closer and clearly separated from normal SI tissue. Lower concentration of glucose, serine and glycine, and increased levels of choline-containing compounds, taurine, lactate and alanine, were found in SI-NETs with more aggressive tumors. Higher abundance of acetate, succinate, choline, phosphocholine, taurine, lactate and aspartate discriminated liver metastases from normal hepatic parenchyma. Higher levels of alanine, ethanolamine, glycerophosphocholine and glucose was found in hepatic metastases than in primary SI-NETs. The present work gives for the first time a snapshot of the metabolomic characteristics of SI-NETs, suggesting the existence of complex metabolic reality, maybe characteristic of different tumor evolution.

Metabolomic profiling highlights the metabolic bases of acute-on-chronic and post-hepatectomy liver failure

BACKGROUND

Posthepatectomy liver failure (PHLF) is the main limitation to extending liver resection but its pathophysiology is not yet fully understood. The aim of the study was to describe the metabolic adaptations that occur with PHLF.

METHODS

A retrospective study of 82 patients using nuclear magnetic resonance metabolomics to identify and quantify intra-hepatic metabolites was performed. The metabolite levels were compared using metabolic network analysis ADEMA between fatal PHLF (FLF) and non fatal PHLF and according to PHLF/ACLF grading.

RESULTS

Metabolomic profiles were significantly different between patients presenting FLF and non FLF or grade 3 ACLF versus < grade 3 ACLF. In the patients undergoing hepatectomy, valine, alanine and glycerophosphocholine were identified as powerful biomarkers to predict FLF (AUROC 0.806, 0.802 and 0.856 respectively). Network analysis showed an activation of aerobic glycolysis with glutaminolysis as observed in highly proliferating systems. Inversely, ACLF3 showed deprivation of glucose and lactate compared to lower ACLF grade.

CONCLUSION

Clinical andbiological severity of ACLF and PHLF correlate with specific metabolic adaptations. Metabolomics can predict fatal liver failure after hepatectomy and underline significant differences in the metabolic patterns of ACLF and PHLF.

Hercules: a profile HMM-based hybrid error correction algorithm for long reads

Choosing whether to use second or third generation sequencing platforms can lead to trade-offs between accuracy and read length. Several types of studies require long and accurate reads. In such cases researchers often combine both technologies and the erroneous long reads are corrected using the short reads. Current approaches rely on various graph or alignment based techniques and do not take the error profile of the underlying technology into account. Efficient machine learning algorithms that address these shortcomings have the potential to achieve more accurate integration of these two technologies. We propose Hercules, the first machine learning-based long read error correction algorithm. Hercules models every long read as a profile Hidden Markov Model with respect to the underlying platform’s error profile. The algorithm learns a posterior transition/emission probability distribution for each long read to correct errors in these reads. We show on two DNA-seq BAC clones (CH17-157L1 and CH17-227A2) that Hercules-corrected reads have the highest mapping rate among all competing algorithms and have the highest accuracy when the breadth of coverage is high. On a large human CHM1 cell line WGS data set, Hercules is one of the few scalable algorithms; and among those, it achieves the highest accuracy.

An integrated genomic and metabolomic approach for defining survival time in adult oligodendrogliomas patients

INTRODUCTION

The identification of frequent acquired mutations shows that patients with oligodendrogliomas have divergent biology with differing prognoses regardless of histological classification. A better understanding of molecular features as well as their metabolic pathways is essential.

OBJECTIVES

The aim of this study was to examine the relationship between the tumor metabolome, six genomic aberrations (isocitrate dehydrogenase1 [IDH1] mutation, 1p/19q codeletion, tumor protein p53 [TP53] mutation, O⁶-methylguanin-DNA methyltransferase [MGMT] promoter methylation, epidermal growth factor receptor [EGFR] amplification, phosphate and tensin homolog [PTEN] methylation), and the patients' survival time.

METHODS

We applied ¹H high-resolution magic-angle spinning (HRMAS) nuclear magnetic resonance (NMR) spectroscopy to 72 resected oligodendrogliomas.

RESULTS

The presence of IDH1, TP53, 1p19q codeletion, MGMT promoter methylation reduced the relative risk of death, whereas PTEN methylation and EGFR amplification were associated with poor prognosis. Increased concentration of 2-hydroxyglutarate (2HG), N-acetyl-aspartate (NAA), myo-inositol and the glycerophosphocholine/phosphocholine (GPC/PC) ratio were good prognostic factors. Increasing the concentration of serine, glycine, glutamate and alanine led to an increased relative risk of death.

CONCLUSION

HRMAS NMR spectroscopy provides accurate information on the metabolomics of oligodendrogliomas, making it possible to find new biomarkers indicative of survival. It enables rapid characterization of intact tissue and could be used as an intraoperative method.

A Statistical Framework for Mapping Risk Genes from De Novo Mutations in Whole-Genome-Sequencing Studies

Analysis of de novo mutations (DNMs) from sequencing data of nuclear families has identified risk genes for many complex diseases, including multiple neurodevelopmental and psychiatric disorders. Most of these efforts have focused on mutations in protein-coding sequences. Evidence from genome-wide association studies (GWASs) strongly suggests that variants important to human diseases often lie in non-coding regions. Extending DNM-based approaches to non-coding sequences is challenging, however, because the functional significance of non-coding mutations is difficult to predict. We propose a statistical framework for analyzing DNMs from whole-genome sequencing (WGS) data. This method, TADA-Annotations (TADA-A), is a major advance of the TADA method we developed earlier for DNM analysis in coding regions. TADA-A is able to incorporate many functional annotations such as conservation and enhancer marks, to learn from data which annotations are informative of pathogenic mutations, and to combine both coding and non-coding mutations at the gene level to detect risk genes. It also supports meta-analysis of multiple DNM studies, while adjusting for study-specific technical effects. We applied TADA-A to WGS data of ∼300 autism-affected family trios across five studies and discovered several autism risk genes. The software is freely available for all research uses.

Characterization of the transcriptional and metabolic responses of pediatric high grade gliomas to mTOR-HIF-1α axis inhibition

Pediatric high grade glioma (pHGGs), including sus-tentorial and diffuse intrinsic pontine gliomas, are known to have a very dismal prognosis. For instance, even an increased knowledge on molecular biology driving this brain tumor entity, there is no treatment able to cure those patients. Therefore, we were focusing on a translational pathway able to increase the cell resistance to treatment and to reprogram metabolically tumor cells, which are, then, adapting easily to a hypoxic microenvironment. To establish, the crucial role of the hypoxic pathways in pHGGs, we, first, assessed their protein and transcriptomic deregulations in a pediatric cohort of pHGGs and in pHGG's cell lines, cultured in both normoxic and hypoxic conditions. Secondly, based on the concept of a bi-therapy targeting in pHGGs mTORC1 (rapamycin) and HIF-1α (irinotecan), we hypothesized that the balanced expressions between RAS/ERK, PI3K/AKT and HIF-1α/HIF-2α/MYC proteins or genes may provide a modulation of the cell response to this double targeting. Finally, we could evidence three protein, genomic and metabolomic profiles of response to rapamycin combined with irinotecan. The pattern of highly sensitive cells to mTOR/HIF-1α targeting was linked to a MYC/ERK/HIF-1α over-expression and the cell resistance to a major hyper-expression of HIF-2α.

Metabolomics approaches in pancreatic adenocarcinoma: tumor metabolism profiling predicts clinical outcome of patients

BACKGROUND

Pancreatic adenocarcinomas (PAs) have very poor prognoses even when surgery is possible. Currently, there are no tissular biomarkers to predict long-term survival in patients with PA. The aims of this study were to (1) describe the metabolome of pancreatic parenchyma (PP) and PA, (2) determine the impact of neoadjuvant chemotherapy on PP and PA, and (3) find tissue metabolic biomarkers associated with long-term survivors, using metabolomics analysis.

METHODS

1H high-resolution magic angle spinning (HRMAS) nuclear magnetic resonance (NMR) spectroscopy using intact tissues was applied to analyze metabolites in PP tissue samples (n = 17) and intact tumor samples (n = 106), obtained from 106 patients undergoing surgical resection for PA.

RESULTS

An orthogonal partial least square-discriminant analysis (OPLS-DA) showed a clear distinction between PP and PA. Higher concentrations of myo-inositol and glycerol were shown in PP, whereas higher levels of glucose, ascorbate, ethanolamine, lactate, and taurine were revealed in PA. Among those metabolites, one of them was particularly obvious in the distinction between long-term and short-term survivors. A high ethanolamine level was associated with worse survival. The impact of neoadjuvant chemotherapy was higher on PA than on PP.

CONCLUSIONS

This study shows that HRMAS NMR spectroscopy using intact tissue provides important and solid information in the characterization of PA. Metabolomics profiling can also predict long-term survival: the assessment of ethanolamine concentration can be clinically relevant as a single metabolic biomarker. This information can be obtained in 20 min, during surgery, to distinguish long-term from short-term survival.

High-resolution magic angle spinning (1)H nuclear magnetic resonance spectroscopy metabolomics of hyperfunctioning parathyroid glands

BACKGROUND

Primary hyperparathyroidism (PHPT) may be related to a single gland disease or multiglandular disease, which requires specific treatments. At present, an operation is the only curative treatment for PHPT. Currently, there are no biomarkers available to identify these 2 entities (single vs. multiple gland disease). The aims of the present study were to compare (1) the tissue metabolomics profiles between PHPT and renal hyperparathyroidism (secondary and tertiary) and (2) single gland disease with multiglandular disease in PHPT using metabolomics analysis.

METHODS

The method used was (1)H high-resolution magic angle spinning nuclear magnetic resonance spectroscopy. Forty-three samples from 32 patients suffering from hyperparathyroidism were included in this study.

RESULTS

Significant differences in the metabolomics profile were assessed according to PHPT and renal hyperparathyroidism. A bicomponent orthogonal partial least square-discriminant analysis showed a clear distinction between PHPT and renal hyperparathyroidism (R(2)Y = 0.85, Q(2) = 0.63). Interestingly, the model also distinguished single gland disease from multiglandular disease (R(2)Y = 0.96, Q(2) = 0.55). A network analysis was also performed using the Algorithm to Determine Expected Metabolite Level Alterations Using Mutual Information (ADEMA). Single gland disease was accurately predicted by ADEMA and was associated with higher levels of phosphorylcholine, choline, glycerophosphocholine, fumarate, succinate, lactate, glucose, glutamine, and ascorbate compared with multiglandular disease.

CONCLUSION

This study shows for the first time that (1)H high-resolution magic angle spinning nuclear magnetic resonance spectroscopy is a reliable and fast technique to distinguish single gland disease from multiglandular disease in patients with PHPT. The potential use of this method as an intraoperative tool requires specific further studies.

De novo ChIP-seq analysis

Methods for the analysis of chromatin immunoprecipitation sequencing (ChIP-seq) data start by aligning the short reads to a reference genome. While often successful, they are not appropriate for cases where a reference genome is not available. Here we develop methods for de novo analysis of ChIP-seq data. Our methods combine de novo assembly with statistical tests enabling motif discovery without the use of a reference genome. We validate the performance of our method using human and mouse data. Analysis of fly data indicates that our method outperforms alignment based methods that utilize closely related species.

MIRA: mutual information-based reporter algorithm for metabolic networks

MOTIVATION

Discovering the transcriptional regulatory architecture of the metabolism has been an important topic to understand the implications of transcriptional fluctuations on metabolism. The reporter algorithm (RA) was proposed to determine the hot spots in metabolic networks, around which transcriptional regulation is focused owing to a disease or a genetic perturbation. Using a z-score-based scoring scheme, RA calculates the average statistical change in the expression levels of genes that are neighbors to a target metabolite in the metabolic network. The RA approach has been used in numerous studies to analyze cellular responses to the downstream genetic changes. In this article, we propose a mutual information-based multivariate reporter algorithm (MIRA) with the goal of eliminating the following problems in detecting reporter metabolites: (i) conventional statistical methods suffer from small sample sizes, (ii) as z-score ranges from minus to plus infinity, calculating average scores can lead to canceling out opposite effects and (iii) analyzing genes one by one, then aggregating results can lead to information loss. MIRA is a multivariate and combinatorial algorithm that calculates the aggregate transcriptional response around a metabolite using mutual information. We show that MIRA's results are biologically sound, empirically significant and more reliable than RA.

RESULTS

We apply MIRA to gene expression analysis of six knockout strains of Escherichia coli and show that MIRA captures the underlying metabolic dynamics of the switch from aerobic to anaerobic respiration. We also apply MIRA to an Autism Spectrum Disorder gene expression dataset. Results indicate that MIRA reports metabolites that highly overlap with recently found metabolic biomarkers in the autism literature. Overall, MIRA is a promising algorithm for detecting metabolic drug targets and understanding the relation between gene expression and metabolic activity.

AVAILABILITY AND IMPLEMENTATION

The code is implemented in C# language using .NET framework. Project is available upon request.

An online model composition tool for system biology models

Background

There are multiple representation formats for Systems Biology computational models, and the Systems Biology Markup Language (SBML) is one of the most widely used. SBML is used to capture, store, and distribute computational models by Systems Biology data sources (e.g., the BioModels Database) and researchers. Therefore, there is a need for all-in-one web-based solutions that support advance SBML functionalities such as uploading, editing, composing, visualizing, simulating, querying, and browsing computational models.

Results

We present the design and implementation of the Model Composition Tool (Interface) within the PathCase-SB (PathCase Systems Biology) web portal. The tool helps users compose systems biology models to facilitate the complex process of merging systems biology models. We also present three tools that support the model composition tool, namely, (1) Model Simulation Interface that generates a visual plot of the simulation according to user’s input, (2) iModel Tool as a platform for users to upload their own models to compose, and (3) SimCom Tool that provides a side by side comparison of models being composed in the same pathway. Finally, we provide a web site that hosts BioModels Database models and a separate web site that hosts SBML Test Suite models.

Conclusions

Model composition tool (and the other three tools) can be used with little or no knowledge of the SBML document structure. For this reason, students or anyone who wants to learn about systems biology will benefit from the described functionalities. SBML Test Suite models will be a nice starting point for beginners. And, for more advanced purposes, users will able to access and employ models of the BioModels Database as well.

ADEMA: an algorithm to determine expected metabolite level alterations using mutual information

Metabolomics is a relatively new “omics” platform, which analyzes a discrete set of metabolites detected in bio-fluids or tissue samples of organisms. It has been used in a diverse array of studies to detect biomarkers and to determine activity rates for pathways based on changes due to disease or drugs. Recent improvements in analytical methodology and large sample throughput allow for creation of large datasets of metabolites that reflect changes in metabolic dynamics due to disease or a perturbation in the metabolic network. However, current methods of comprehensive analyses of large metabolic datasets (metabolomics) are limited, unlike other “omics” approaches where complex techniques for analyzing coexpression/coregulation of multiple variables are applied. This paper discusses the shortcomings of current metabolomics data analysis techniques, and proposes a new multivariate technique (ADEMA) based on mutual information to identify expected metabolite level changes with respect to a specific condition. We show that ADEMA better predicts De Novo Lipogenesis pathway metabolite level changes in samples with Cystic Fibrosis (CF) than prediction based on the significance of individual metabolite level changes. We also applied ADEMA's classification scheme on three different cohorts of CF and wildtype mice. ADEMA was able to predict whether an unknown mouse has a CF or a wildtype genotype with 1.0, 0.84, and 0.9 accuracy for each respective dataset. ADEMA results had up to 31% higher accuracy as compared to other classification algorithms. In conclusion, ADEMA advances the state-of-the-art in metabolomics analysis, by providing accurate and interpretable classification results.

Metabolomics is an experimental approach that analyzes differences in metabolite levels detected in experimental samples. It has been used in the literature to understand the changes in metabolism with respect to diseases or drugs. Unlike transcriptomics or proteomics, which analyze gene and protein expression levels respectively, the techniques that consider co-regulation of multiple metabolites are quite limited. In this paper, we propose a novel technique, called ADEMA, which computes the expected level changes for each metabolite with respect to a given condition. ADEMA considers multiple metabolites at the same time and is mutual information (MI)-based. We show that ADEMA predicts metabolite level changes for young mice with Cystic Fibrosis (CF) better than significance testing that considers one metabolite at a time. Using three different datasets that contain CF and wild-type (WT) mice, we show that ADEMA can classify an individual as being CF or WT based on the metabolic profiles (with 1.0, 0.84, and 0.9 accuracy, respectively). Compared to other well-known classification algorithms, ADEMA's accuracy is higher by up to 31%.

Observation conflict resolution in steady-state metabolic network dynamics analysis

Steady state metabolic network dynamics analysis (SMDA) is a recently proposed computational metabolomics tool that (i) captures a metabolic network and its rules via a metabolic network database, (ii) mimics the reasoning of a biochemist, given a set of metabolic observations, and (iii) locates efficiently all possible metabolic activation/inactivation (flux) alternatives. However, a number of factors may cause the SMDA algorithm to eliminate feasible flux scenarios. These factors include (i) inherent error margins in observations (measurements), (ii) lack of knowledge to classify measurements as normal versus abnormal, and (iii) choosing a highly constrained metabolic subnetwork to query against. In this work, we first present and formalize these obstacles. Then, we propose techniques to eliminate them and present an experimental evaluation of our proposed techniques.

PathCase-SB: integrating data sources and providing tools for systems biology research

BACKGROUND

Integration of metabolic pathways resources and metabolic network models, and deploying new tools on the integrated platform can help perform more effective and more efficient systems biology research on understanding the regulation of metabolic networks. Therefore, the tasks of (a) integrating under a single database environment regulatory metabolic networks and existing models, and (b) building tools to help with modeling and analysis are desirable and intellectually challenging computational tasks.

RESULTS

PathCase Systems Biology (PathCase-SB) is built and released. This paper describes PathCase-SB user interfaces developed to date. The current PathCase-SB system provides a database-enabled framework and web-based computational tools towards facilitating the development of kinetic models for biological systems. PathCase-SB aims to integrate systems biology models data and metabolic network data of selected biological data sources on the web (currently, BioModels Database and KEGG, respectively), and to provide more powerful and/or new capabilities via the new web-based integrative framework.

CONCLUSIONS

Each of the current four PathCase-SB interfaces, namely, Browser, Visualization, Querying, and Simulation interfaces, have expanded and new capabilities as compared with the original data sources. PathCase-SB is already available on the web and being used by researchers across the globe.

PathCase-SB architecture and database design

Background

Integration of metabolic pathways resources and regulatory metabolic network models, and deploying new tools on the integrated platform can help perform more effective and more efficient systems biology research on understanding the regulation in metabolic networks. Therefore, the tasks of (a) integrating under a single database environment regulatory metabolic networks and existing models, and (b) building tools to help with modeling and analysis are desirable and intellectually challenging computational tasks.

Description

PathCase Systems Biology (PathCase-SB) is built and released. The PathCase-SB database provides data and API for multiple user interfaces and software tools. The current PathCase-SB system provides a database-enabled framework and web-based computational tools towards facilitating the development of kinetic models for biological systems. PathCase-SB aims to integrate data of selected biological data sources on the web (currently, BioModels database and KEGG), and to provide more powerful and/or new capabilities via the new web-based integrative framework. This paper describes architecture and database design issues encountered in PathCase-SB's design and implementation, and presents the current design of PathCase-SB's architecture and database.

Conclusions

PathCase-SB architecture and database provide a highly extensible and scalable environment with easy and fast (real-time) access to the data in the database. PathCase-SB itself is already being used by researchers across the world.