Gene expression data can be used to predict synergistic drug combinations for cancer treatment. The author developed three algorithms to do this - regularized bilinear regression, up-down gene analysis regression, and a neighborhood predictor. Performance improved with each subsequent model as they incorporated more biological context. The neighborhood predictor performed best by looking at similar drug combinations and cell lines based on gene expression similarity. This suggests gene expression is key for predicting synergy and combining drugs targeted at multiple pathways may overcome drug resistance in cancer. More gene expression data on more cell lines and drug perturbations could further improve predictions of synergistic combinations.
S1: Introduces a study titled "Anticancer Thiazolidinones Design: Mining of 60-Cell Lines Experimental Data" which aims to extract valuable knowledge from experimental anticancer screening results to inform future QSAR studies.
S2: Describes the two-stage in vitro screening process conducted by the National Cancer Institute on 60 human tumor cell lines to evaluate compound growth inhibition.
S3: Outlines problems to be addressed: whether same-dose results are statistically similar enough to combine, how to define active vs inactive compounds, and whether different antitumor mechanisms are present.
The document discusses mixed models, which contain both fixed and random effects. Fixed effects have all possible levels included in the study, while random effects are a random sample from the total population. The mixed model is represented as Y = Xβ + Zγ + ε, where β are fixed effects, X are fixed effect variables, Z are random effects, γ are random effect parameters, and ε is the error term. Mixed models can model both fixed and random effects, account for correlation in errors, and handle missing data. They provide correct standard errors compared to general linear models (GLMs). Model fitting involves likelihood ratio tests and information criteria to select the best fitting model.
Target Based Drug Combination SelectionRajeev Gangal
The presentation outlines an algorithm to identify combinations of drug for a given therapeutic endpoint.
The objective is to target different stages of the disease pathway.
Chemoinformatics as employed by Ontomine, a US patent pending algorithm is employed for the same.
Mixed Models: How to Effectively Account for Inbreeding and Population Struct...Golden Helix Inc
Population structure and inbreeding can confound results from a standard genome-wide association test. Accounting for the random effect of relatedness can lead to lower false discovery rates and identify the causative markers without over-correcting and dampening the true signal.
This presentation will review four different methods of analyzing genotype data while accounting for random effects of relatedness. Methods include PCA analysis with Linear Regression, GBLUP, EMMAX, and MLMM. Comparisons will be made using data from the Sheep HapMap project and a simulated phenotype. After presenting the various methods, we will discuss how these results can be obtained using Golden Helix SNP & Variation Suite (SVS) software and how SVS can be used to compare and contrast the results.
- When two drugs exhibit a constant potency ratio in producing an effect, their combination will produce a linear additive isobole that can distinguish synergistic and antagonistic interactions.
- The additive isobole is based on the historical concept of dose equivalence.
- When one drug is a partial agonist, the additive isobole calculated using dose equivalence requires large enough doses to produce the maximum effect, resulting in non-linear isoboles.
A comparative study of covariance selection models for the inference of gene ...Roberto Anglani
This study compares three methods for estimating gene regulatory networks from gene expression data: 1) a pseudoinverse method (PINV) that estimates the precision matrix using the Moore-Penrose pseudoinverse of the sample covariance matrix, 2) a regularized least squares method (RCM) that estimates partial correlations from regression residuals, and 3) a regularized log-likelihood method ('2C) that maximizes a penalized log-likelihood function to estimate the precision matrix. Extensive simulations show that the '2C method has the most predictive partial correlations and highest sensitivity for inferring conditional dependencies. Application to real datasets provides biological insights into gene pathways in Arabidopsis and human cells.
The Indian Dental Academy is the Leader in continuing dental education , training dentists in all aspects of dentistry and offering a wide range of dental certified courses in different formats.
The documents discuss graded dose response curves which depict the relationship between drug dose and magnitude of effect. The potency of a drug refers to the amount needed to produce a response, which can be measured by the dose required to produce half the maximum effect. Effectiveness considers both response and safety, which is measured by the therapeutic index - the ratio of lethal to effective doses. Quantal dose response curves plot the percentage of a population responding at given doses and are useful in determining therapeutic indices.
S1: Introduces a study titled "Anticancer Thiazolidinones Design: Mining of 60-Cell Lines Experimental Data" which aims to extract valuable knowledge from experimental anticancer screening results to inform future QSAR studies.
S2: Describes the two-stage in vitro screening process conducted by the National Cancer Institute on 60 human tumor cell lines to evaluate compound growth inhibition.
S3: Outlines problems to be addressed: whether same-dose results are statistically similar enough to combine, how to define active vs inactive compounds, and whether different antitumor mechanisms are present.
The document discusses mixed models, which contain both fixed and random effects. Fixed effects have all possible levels included in the study, while random effects are a random sample from the total population. The mixed model is represented as Y = Xβ + Zγ + ε, where β are fixed effects, X are fixed effect variables, Z are random effects, γ are random effect parameters, and ε is the error term. Mixed models can model both fixed and random effects, account for correlation in errors, and handle missing data. They provide correct standard errors compared to general linear models (GLMs). Model fitting involves likelihood ratio tests and information criteria to select the best fitting model.
Target Based Drug Combination SelectionRajeev Gangal
The presentation outlines an algorithm to identify combinations of drug for a given therapeutic endpoint.
The objective is to target different stages of the disease pathway.
Chemoinformatics as employed by Ontomine, a US patent pending algorithm is employed for the same.
Mixed Models: How to Effectively Account for Inbreeding and Population Struct...Golden Helix Inc
Population structure and inbreeding can confound results from a standard genome-wide association test. Accounting for the random effect of relatedness can lead to lower false discovery rates and identify the causative markers without over-correcting and dampening the true signal.
This presentation will review four different methods of analyzing genotype data while accounting for random effects of relatedness. Methods include PCA analysis with Linear Regression, GBLUP, EMMAX, and MLMM. Comparisons will be made using data from the Sheep HapMap project and a simulated phenotype. After presenting the various methods, we will discuss how these results can be obtained using Golden Helix SNP & Variation Suite (SVS) software and how SVS can be used to compare and contrast the results.
- When two drugs exhibit a constant potency ratio in producing an effect, their combination will produce a linear additive isobole that can distinguish synergistic and antagonistic interactions.
- The additive isobole is based on the historical concept of dose equivalence.
- When one drug is a partial agonist, the additive isobole calculated using dose equivalence requires large enough doses to produce the maximum effect, resulting in non-linear isoboles.
A comparative study of covariance selection models for the inference of gene ...Roberto Anglani
This study compares three methods for estimating gene regulatory networks from gene expression data: 1) a pseudoinverse method (PINV) that estimates the precision matrix using the Moore-Penrose pseudoinverse of the sample covariance matrix, 2) a regularized least squares method (RCM) that estimates partial correlations from regression residuals, and 3) a regularized log-likelihood method ('2C) that maximizes a penalized log-likelihood function to estimate the precision matrix. Extensive simulations show that the '2C method has the most predictive partial correlations and highest sensitivity for inferring conditional dependencies. Application to real datasets provides biological insights into gene pathways in Arabidopsis and human cells.
The Indian Dental Academy is the Leader in continuing dental education , training dentists in all aspects of dentistry and offering a wide range of dental certified courses in different formats.
The documents discuss graded dose response curves which depict the relationship between drug dose and magnitude of effect. The potency of a drug refers to the amount needed to produce a response, which can be measured by the dose required to produce half the maximum effect. Effectiveness considers both response and safety, which is measured by the therapeutic index - the ratio of lethal to effective doses. Quantal dose response curves plot the percentage of a population responding at given doses and are useful in determining therapeutic indices.
- The document proposes a multi-view stacking ensemble method for drug-target interaction (DTI) prediction that combines predictions from multiple machine learning models trained on different drug and target feature view combinations.
- It generates 126 view combination datasets from 14 drug views and 9 target views, then trains extra trees, random forest, and XGBoost classifiers on each view combination. Predictions from these base models are then combined using a stacking ensemble with an extra trees meta-learner.
- The method is shown to outperform single models and voting ensembles, and calibration of the meta-learner and use of local imbalance measures provide further improvements to predictive performance on DTI prediction tasks.
The document summarizes a study that characterized copy number alterations and mutations in primary breast cancer samples. It found that the most frequent amplified regions were on chromosomes 1q, 8p12, 8q24, 11q13, and others. These amplified regions formed an "amplicome" of 30 regions containing 1,747 genes. While individual amplified regions did not strongly correlate with specific pathways, together they synergistically encoded many processes related to tumorigenesis. The study also analyzed mutations and found they were distributed randomly throughout the genome, in contrast to amplifications. Mutated genes tended to be transcription factors upstream of amplified gene targets. Both amplifications and mutations cooperatively enriched pathways driving breast cancer.
The Principle of Rational Design of Drug Combination and Personalized Therapy...Jianghui Xiong
This document discusses principles of rational drug combination design and personalized therapy based on network pharmacology. It provides several examples:
1) Using gene expression signatures to identify drug combinations that improve drug sensitivity, such as dexamethasone and sirolimus for acute lymphoblastic leukemia.
2) Designing combinations based on synthetic lethal screens, such as identifying genes that sensitize cancer cells to epidermal growth factor receptor inhibitors.
3) A strategy for personalized cancer therapy based on identifying genes with synthetic lethal interactions with oncogenes like KRAS, and using these genes as therapy targets depending on a patient's mutation status.
4) A concept called "synergistic outcome determination" to model
This document describes the development of gene expression signatures that can predict sensitivity to various chemotherapeutic drugs using microarray data from cancer cell lines.
1) Signatures were developed for several drugs including docetaxel, topotecan, adriamycin, etoposide, 5-fluorouracil, paclitaxel, and cyclophosphamide that could accurately predict drug sensitivity in independent cancer cell line datasets.
2) These signatures were also shown to predict clinical response to the drugs in human patients, including predicting response to docetaxel in breast cancer and ovarian cancer with over 85% accuracy.
3) The signatures were specific to each individual drug and could predict response to multid
This document discusses using hierarchical clustering algorithms to mine pharmacogenomic data from a clinical data warehouse containing both genomic and clinical trial data. It describes hierarchical clustering as an approach to analyze gene expression microarray data by grouping single profiles into clusters in a tree structure. Six main hierarchical clustering algorithms are outlined that differ in how they calculate distances between clusters. The goal of pharmacogenomic data mining is to discover knowledge that can identify the most effective and least toxic drugs for individuals based on their genetic makeup and disease.
Comparing Genetic Evolutionary Algorithms on Three Enzymes of HIV-1: Integras...CSCJournals
In this work, we utilized Quantitative Structure-Activity Relationship (QSAR) techniques to develop predictive models for inhibitors of the HIV-1 enzymes Integrase, HIV-Protease, and Reverse Transcriptase. Each predictive model was composed of quantitative drug characteristics that were selected by genetic evolutionary algorithms, such as Genetic Algorithm (GE), Differential Evolutionary Algorithm (DE), Binary Particle Swarm Optimization (BPSO), and Differential Evolution with Binary Particle Swarm Optimization (DE-BPSO). After characteristic selection, each model was tested with machine-learning algorithms such as Multiple Linear Regression (MLR), Support Vector Machine (SVM), and Multi-Layer Perceptron neural networks (MLP/ANN). We found that a combination of DE-BPSO combined with Multi-Layer Perceptron produced the most accurate predictive models as measured by R2, the statistical measure of proportion of variance in prediction values, and root-mean-square-error (RMSE) of prediction values compared to observed values. As for the models themselves: the best predictors for Integrase inhibitor included mass-weighted centred Broto-Moreau autocorrelation values, Moran autocorrelations, and eigenvalues of Burden matrices weighted by I-states; the best predictors for HIV-Protease inhibitors included the second Zagreb index value, the normalized spectral positive sum from Laplace matrix, and the connectivity-like index of order 0 from edge adjacency mat; and the best predictors for Reverse Transcriptase inhibitors included the number of hydrogen atoms, the molecular path count of order 7, the centred Broto-Moreau autocorrelation of lag 2 weighted by Sanderson electronegativity, the P_VSA-like on ionization potential, and the frequency of C – N bonds at topological distance 3.
Rational drug design involves developing compounds that target specific biomolecules involved in disease processes through protein-protein or protein-nucleic acid interactions. Protein targets can be identified through techniques like X-ray crystallography and NMR. Computational tools and global gene expression analysis help increase the efficiency and cost-effectiveness of the drug design process by aiding in structure-guided approaches and target identification. Drug design can involve developing ligands for targets with known structures or developing ligands with predefined properties for unknown targets identified through gene expression data. Combination therapies and overcoming toxic side effects are important challenges in developing improved anti-cancer drugs.
Optimal drug prediction from personal genomics profilesAarathi Anil
The document describes a proposed system for optimally predicting drugs for cancer patients based on their personal genomics profiles. The system uses drug sensitivity data from cancer cell lines and gene expression profiles to calculate similarity scores between drugs and between patients. These scores are combined to assign a drug-patient effect score, predicting the most sensitive drugs for individual patients. Validation tests on independent datasets showed the system significantly improves effective drug selection over random chance. The system could help select optimal personalized therapies for cancer patients based on their genomic information.
Identified the likelihood of success for treatment of five cancers of interest by comparing novel drug combinations treated cell line gene signatures (predicted by bioinformatic analysis), with disease gene signatures (calculated by fitting linear model on gene expression data).
A method for mining infrequent causal associations and its application in fin...IEEEFINALYEARPROJECTS
To Get any Project for CSE, IT ECE, EEE Contact Me @ 09849539085, 09966235788 or mail us - ieeefinalsemprojects@gmail.co¬m-Visit Our Website: www.finalyearprojects.org
This document discusses how drug analytics based on manually extracted semantic relationships in Embase can be useful for drug development, repurposing, and safety. It describes how relationships between drugs, diseases, and adverse reactions that are manually indexed can provide valuable information for drug repurposing, development, and safety. Specific examples are provided to show how the semantic relationships can guide drug repositioning strategies, investigate new combination drugs, identify drug-drug interactions, collect drug comparison data, and help improve risk management.
1. Statistical analysis of big data sets from microarrays and RNA-seq is used to identify differentially expressed genes. Heat maps and volcano plots are commonly used to visualize the data.
2. Gene ontology, gene set enrichment analysis, and transcription factor analysis are used to analyze lists of genes and identify biological processes, pathways, and regulatory relationships.
3. Networks can be constructed by integrating gene lists with protein-protein and gene regulatory interaction databases to build signaling, regulatory, and interaction networks for further analysis.
This document summarizes statistical methods for analyzing cDNA microarray data, including data preprocessing, normalization techniques, and statistical tests. It discusses alignment, background calculation, data transformation, normalization methods like global normalization, housekeeping gene normalization, and intensity-dependent normalization. Statistical tests covered include t-tests, multiple testing adjustments, permutation tests, and significance analysis of microarrays (SAM). The document concludes that no single method is best and different data may require trying different analytical approaches.
Data analysis and Visualisation Techniques for Compound Combination ModellingRichard Lewis
A talk given at the EBI for the Cambridge Cheminformatics Network Meeting on the 25/11/2015, introducing compound combination modelling analysis and visualisation techniques.
Network analysis of cancer metabolism: A novel route to precision medicineVarshit Dusad
This document discusses using network analysis and mass flow graphs to analyze cancer cell metabolism. It assesses different published genome-scale metabolic models of cancer and determines that PRIME models are best suited for applying mass flow graph analysis. Constraint-based analysis is performed on PRIME models to simulate metabolic conditions and genetic perturbations. Centrality analysis using PageRank reveals changes in network structure under different conditions but does not fully support the centrality-lethality hypothesis regarding essential reactions. Future work is needed to better integrate omics data and identify centrality measures that correlate with biological importance.
Selection of genes to include in genomic studies of disease
remains a difficult task. Current methods rely on expert opinion
or manual search engine use. With these methods, the
process and result are neither repeatable nor scalable. To
remedy this situation, we created the Informative Genetic
Content (IGC) system, which enables the algorithmic selection
of genes for inclusion in such studies, given one or more
diseases to target.
The IGC system stands on three components: a database
associating diseases with genes and other diseases, an
algorithm to rank the genes under consideration for inclusion in
a panel, and a module that clusters genes by families of
diseases. The first component, the database, maps diseases
to associated genes and scores each of these mappings
according to the strength of the relationship. The database also
maps diseases to other diseases, such that groups of diseases
or hierarchical relationships between diseases can be
identified. The second component enables the ranking of
candidate genes when multiple diseases are of interest. The
algorithm accounts for the common situation where two or
more diseases are associated with the same gene with varying
strengths of association, weighting and combining the scores
across the diseases associated with each gene. The final
component, the gene clustering module, groups genes by
pathogenic pathways, should the user want to consider
targeting a broader family of diseases affected by a closely
related set of genes.
We validated the IGC system through comparisons of our
automated gene selections with expertly curated gene panel
designs. We found a high degree of overlap between the IGC’s
gene selection and the gene lists chosen by experts,
supporting the viability of our system.
Together with the scalability and repeatability enabled by its
automation, the IGC system greatly improves the gene panel
selection process and therefore advances targeted genomic
studies.
Rasamanikya is a excellent preparation in the field of Rasashastra, it is used in various Kushtha Roga, Shwasa, Vicharchika, Bhagandara, Vatarakta, and Phiranga Roga. In this article Preparation& Comparative analytical profile for both Formulationon i.e Rasamanikya prepared by Kushmanda swarasa & Churnodhaka Shodita Haratala. The study aims to provide insights into the comparative efficacy and analytical aspects of these formulations for enhanced therapeutic outcomes.
More Related Content
Similar to Novel Methodology for Predicting Synergistic Cancer Drug Pairs Slides
- The document proposes a multi-view stacking ensemble method for drug-target interaction (DTI) prediction that combines predictions from multiple machine learning models trained on different drug and target feature view combinations.
- It generates 126 view combination datasets from 14 drug views and 9 target views, then trains extra trees, random forest, and XGBoost classifiers on each view combination. Predictions from these base models are then combined using a stacking ensemble with an extra trees meta-learner.
- The method is shown to outperform single models and voting ensembles, and calibration of the meta-learner and use of local imbalance measures provide further improvements to predictive performance on DTI prediction tasks.
The document summarizes a study that characterized copy number alterations and mutations in primary breast cancer samples. It found that the most frequent amplified regions were on chromosomes 1q, 8p12, 8q24, 11q13, and others. These amplified regions formed an "amplicome" of 30 regions containing 1,747 genes. While individual amplified regions did not strongly correlate with specific pathways, together they synergistically encoded many processes related to tumorigenesis. The study also analyzed mutations and found they were distributed randomly throughout the genome, in contrast to amplifications. Mutated genes tended to be transcription factors upstream of amplified gene targets. Both amplifications and mutations cooperatively enriched pathways driving breast cancer.
The Principle of Rational Design of Drug Combination and Personalized Therapy...Jianghui Xiong
This document discusses principles of rational drug combination design and personalized therapy based on network pharmacology. It provides several examples:
1) Using gene expression signatures to identify drug combinations that improve drug sensitivity, such as dexamethasone and sirolimus for acute lymphoblastic leukemia.
2) Designing combinations based on synthetic lethal screens, such as identifying genes that sensitize cancer cells to epidermal growth factor receptor inhibitors.
3) A strategy for personalized cancer therapy based on identifying genes with synthetic lethal interactions with oncogenes like KRAS, and using these genes as therapy targets depending on a patient's mutation status.
4) A concept called "synergistic outcome determination" to model
This document describes the development of gene expression signatures that can predict sensitivity to various chemotherapeutic drugs using microarray data from cancer cell lines.
1) Signatures were developed for several drugs including docetaxel, topotecan, adriamycin, etoposide, 5-fluorouracil, paclitaxel, and cyclophosphamide that could accurately predict drug sensitivity in independent cancer cell line datasets.
2) These signatures were also shown to predict clinical response to the drugs in human patients, including predicting response to docetaxel in breast cancer and ovarian cancer with over 85% accuracy.
3) The signatures were specific to each individual drug and could predict response to multid
This document discusses using hierarchical clustering algorithms to mine pharmacogenomic data from a clinical data warehouse containing both genomic and clinical trial data. It describes hierarchical clustering as an approach to analyze gene expression microarray data by grouping single profiles into clusters in a tree structure. Six main hierarchical clustering algorithms are outlined that differ in how they calculate distances between clusters. The goal of pharmacogenomic data mining is to discover knowledge that can identify the most effective and least toxic drugs for individuals based on their genetic makeup and disease.
Comparing Genetic Evolutionary Algorithms on Three Enzymes of HIV-1: Integras...CSCJournals
In this work, we utilized Quantitative Structure-Activity Relationship (QSAR) techniques to develop predictive models for inhibitors of the HIV-1 enzymes Integrase, HIV-Protease, and Reverse Transcriptase. Each predictive model was composed of quantitative drug characteristics that were selected by genetic evolutionary algorithms, such as Genetic Algorithm (GE), Differential Evolutionary Algorithm (DE), Binary Particle Swarm Optimization (BPSO), and Differential Evolution with Binary Particle Swarm Optimization (DE-BPSO). After characteristic selection, each model was tested with machine-learning algorithms such as Multiple Linear Regression (MLR), Support Vector Machine (SVM), and Multi-Layer Perceptron neural networks (MLP/ANN). We found that a combination of DE-BPSO combined with Multi-Layer Perceptron produced the most accurate predictive models as measured by R2, the statistical measure of proportion of variance in prediction values, and root-mean-square-error (RMSE) of prediction values compared to observed values. As for the models themselves: the best predictors for Integrase inhibitor included mass-weighted centred Broto-Moreau autocorrelation values, Moran autocorrelations, and eigenvalues of Burden matrices weighted by I-states; the best predictors for HIV-Protease inhibitors included the second Zagreb index value, the normalized spectral positive sum from Laplace matrix, and the connectivity-like index of order 0 from edge adjacency mat; and the best predictors for Reverse Transcriptase inhibitors included the number of hydrogen atoms, the molecular path count of order 7, the centred Broto-Moreau autocorrelation of lag 2 weighted by Sanderson electronegativity, the P_VSA-like on ionization potential, and the frequency of C – N bonds at topological distance 3.
Rational drug design involves developing compounds that target specific biomolecules involved in disease processes through protein-protein or protein-nucleic acid interactions. Protein targets can be identified through techniques like X-ray crystallography and NMR. Computational tools and global gene expression analysis help increase the efficiency and cost-effectiveness of the drug design process by aiding in structure-guided approaches and target identification. Drug design can involve developing ligands for targets with known structures or developing ligands with predefined properties for unknown targets identified through gene expression data. Combination therapies and overcoming toxic side effects are important challenges in developing improved anti-cancer drugs.
Optimal drug prediction from personal genomics profilesAarathi Anil
The document describes a proposed system for optimally predicting drugs for cancer patients based on their personal genomics profiles. The system uses drug sensitivity data from cancer cell lines and gene expression profiles to calculate similarity scores between drugs and between patients. These scores are combined to assign a drug-patient effect score, predicting the most sensitive drugs for individual patients. Validation tests on independent datasets showed the system significantly improves effective drug selection over random chance. The system could help select optimal personalized therapies for cancer patients based on their genomic information.
Identified the likelihood of success for treatment of five cancers of interest by comparing novel drug combinations treated cell line gene signatures (predicted by bioinformatic analysis), with disease gene signatures (calculated by fitting linear model on gene expression data).
A method for mining infrequent causal associations and its application in fin...IEEEFINALYEARPROJECTS
To Get any Project for CSE, IT ECE, EEE Contact Me @ 09849539085, 09966235788 or mail us - ieeefinalsemprojects@gmail.co¬m-Visit Our Website: www.finalyearprojects.org
This document discusses how drug analytics based on manually extracted semantic relationships in Embase can be useful for drug development, repurposing, and safety. It describes how relationships between drugs, diseases, and adverse reactions that are manually indexed can provide valuable information for drug repurposing, development, and safety. Specific examples are provided to show how the semantic relationships can guide drug repositioning strategies, investigate new combination drugs, identify drug-drug interactions, collect drug comparison data, and help improve risk management.
1. Statistical analysis of big data sets from microarrays and RNA-seq is used to identify differentially expressed genes. Heat maps and volcano plots are commonly used to visualize the data.
2. Gene ontology, gene set enrichment analysis, and transcription factor analysis are used to analyze lists of genes and identify biological processes, pathways, and regulatory relationships.
3. Networks can be constructed by integrating gene lists with protein-protein and gene regulatory interaction databases to build signaling, regulatory, and interaction networks for further analysis.
This document summarizes statistical methods for analyzing cDNA microarray data, including data preprocessing, normalization techniques, and statistical tests. It discusses alignment, background calculation, data transformation, normalization methods like global normalization, housekeeping gene normalization, and intensity-dependent normalization. Statistical tests covered include t-tests, multiple testing adjustments, permutation tests, and significance analysis of microarrays (SAM). The document concludes that no single method is best and different data may require trying different analytical approaches.
Data analysis and Visualisation Techniques for Compound Combination ModellingRichard Lewis
A talk given at the EBI for the Cambridge Cheminformatics Network Meeting on the 25/11/2015, introducing compound combination modelling analysis and visualisation techniques.
Network analysis of cancer metabolism: A novel route to precision medicineVarshit Dusad
This document discusses using network analysis and mass flow graphs to analyze cancer cell metabolism. It assesses different published genome-scale metabolic models of cancer and determines that PRIME models are best suited for applying mass flow graph analysis. Constraint-based analysis is performed on PRIME models to simulate metabolic conditions and genetic perturbations. Centrality analysis using PageRank reveals changes in network structure under different conditions but does not fully support the centrality-lethality hypothesis regarding essential reactions. Future work is needed to better integrate omics data and identify centrality measures that correlate with biological importance.
Selection of genes to include in genomic studies of disease
remains a difficult task. Current methods rely on expert opinion
or manual search engine use. With these methods, the
process and result are neither repeatable nor scalable. To
remedy this situation, we created the Informative Genetic
Content (IGC) system, which enables the algorithmic selection
of genes for inclusion in such studies, given one or more
diseases to target.
The IGC system stands on three components: a database
associating diseases with genes and other diseases, an
algorithm to rank the genes under consideration for inclusion in
a panel, and a module that clusters genes by families of
diseases. The first component, the database, maps diseases
to associated genes and scores each of these mappings
according to the strength of the relationship. The database also
maps diseases to other diseases, such that groups of diseases
or hierarchical relationships between diseases can be
identified. The second component enables the ranking of
candidate genes when multiple diseases are of interest. The
algorithm accounts for the common situation where two or
more diseases are associated with the same gene with varying
strengths of association, weighting and combining the scores
across the diseases associated with each gene. The final
component, the gene clustering module, groups genes by
pathogenic pathways, should the user want to consider
targeting a broader family of diseases affected by a closely
related set of genes.
We validated the IGC system through comparisons of our
automated gene selections with expertly curated gene panel
designs. We found a high degree of overlap between the IGC’s
gene selection and the gene lists chosen by experts,
supporting the viability of our system.
Together with the scalability and repeatability enabled by its
automation, the IGC system greatly improves the gene panel
selection process and therefore advances targeted genomic
studies.
Rasamanikya is a excellent preparation in the field of Rasashastra, it is used in various Kushtha Roga, Shwasa, Vicharchika, Bhagandara, Vatarakta, and Phiranga Roga. In this article Preparation& Comparative analytical profile for both Formulationon i.e Rasamanikya prepared by Kushmanda swarasa & Churnodhaka Shodita Haratala. The study aims to provide insights into the comparative efficacy and analytical aspects of these formulations for enhanced therapeutic outcomes.
Recomendações da OMS sobre cuidados maternos e neonatais para uma experiência pós-natal positiva.
Em consonância com os ODS – Objetivos do Desenvolvimento Sustentável e a Estratégia Global para a Saúde das Mulheres, Crianças e Adolescentes, e aplicando uma abordagem baseada nos direitos humanos, os esforços de cuidados pós-natais devem expandir-se para além da cobertura e da simples sobrevivência, de modo a incluir cuidados de qualidade.
Estas diretrizes visam melhorar a qualidade dos cuidados pós-natais essenciais e de rotina prestados às mulheres e aos recém-nascidos, com o objetivo final de melhorar a saúde e o bem-estar materno e neonatal.
Uma “experiência pós-natal positiva” é um resultado importante para todas as mulheres que dão à luz e para os seus recém-nascidos, estabelecendo as bases para a melhoria da saúde e do bem-estar a curto e longo prazo. Uma experiência pós-natal positiva é definida como aquela em que as mulheres, pessoas que gestam, os recém-nascidos, os casais, os pais, os cuidadores e as famílias recebem informação consistente, garantia e apoio de profissionais de saúde motivados; e onde um sistema de saúde flexível e com recursos reconheça as necessidades das mulheres e dos bebês e respeite o seu contexto cultural.
Estas diretrizes consolidadas apresentam algumas recomendações novas e já bem fundamentadas sobre cuidados pós-natais de rotina para mulheres e neonatos que recebem cuidados no pós-parto em unidades de saúde ou na comunidade, independentemente dos recursos disponíveis.
É fornecido um conjunto abrangente de recomendações para cuidados durante o período puerperal, com ênfase nos cuidados essenciais que todas as mulheres e recém-nascidos devem receber, e com a devida atenção à qualidade dos cuidados; isto é, a entrega e a experiência do cuidado recebido. Estas diretrizes atualizam e ampliam as recomendações da OMS de 2014 sobre cuidados pós-natais da mãe e do recém-nascido e complementam as atuais diretrizes da OMS sobre a gestão de complicações pós-natais.
O estabelecimento da amamentação e o manejo das principais intercorrências é contemplada.
Recomendamos muito.
Vamos discutir essas recomendações no nosso curso de pós-graduação em Aleitamento no Instituto Ciclos.
Esta publicação só está disponível em inglês até o momento.
Prof. Marcus Renato de Carvalho
www.agostodourado.com
Osteoporosis - Definition , Evaluation and Management .pdfJim Jacob Roy
Osteoporosis is an increasing cause of morbidity among the elderly.
In this document , a brief outline of osteoporosis is given , including the risk factors of osteoporosis fractures , the indications for testing bone mineral density and the management of osteoporosis
ABDOMINAL TRAUMA in pediatrics part one.drhasanrajab
Abdominal trauma in pediatrics refers to injuries or damage to the abdominal organs in children. It can occur due to various causes such as falls, motor vehicle accidents, sports-related injuries, and physical abuse. Children are more vulnerable to abdominal trauma due to their unique anatomical and physiological characteristics. Signs and symptoms include abdominal pain, tenderness, distension, vomiting, and signs of shock. Diagnosis involves physical examination, imaging studies, and laboratory tests. Management depends on the severity and may involve conservative treatment or surgical intervention. Prevention is crucial in reducing the incidence of abdominal trauma in children.
share - Lions, tigers, AI and health misinformation, oh my!.pptxTina Purnat
• Pitfalls and pivots needed to use AI effectively in public health
• Evidence-based strategies to address health misinformation effectively
• Building trust with communities online and offline
• Equipping health professionals to address questions, concerns and health misinformation
• Assessing risk and mitigating harm from adverse health narratives in communities, health workforce and health system
Does Over-Masturbation Contribute to Chronic Prostatitis.pptxwalterHu5
In some case, your chronic prostatitis may be related to over-masturbation. Generally, natural medicine Diuretic and Anti-inflammatory Pill can help mee get a cure.
Novel Methodology for Predicting Synergistic Cancer Drug Pairs Slides
1. Gene Expression as a Key
Component of Predicting
Synergistic Drug Combinations
Megan Yin
2. Background
Limitations of Targeted
Therapies
Cancer is a multigenic disease
Targeted therapies are
ineffective because of
acquired drug resistance
Despite increased government
investment, there is a
decline in drug discovery
Next Steps with Combination
Therapy
Combination therapy rests on
the assumption that effective
drug combinations are known
Want drugs that work together
synergistically
Hard to explore this in vitro due
to the large combinatorial
space
Currently no good
methodologies available to
predict drug synergy
2
Combination Therapy
Can get past limitations of drug
resistance by targeting
multiple pathways at the
same time to kill tumor cells
Uses currently available drugs
only
Decrease drug dosages since
drugs can be more potent in
combination
3. Literature Review
Median Effect
Median-effect
based models
depend on the
linearity of the
median-effect
plot which is not
entirely fulfilled in
the context of
cancer
combinations.
Loewe Additivity
Loewe additivity
assumes that a
dose-response
curve is known
but dose-response
curves are not
known due to
noise or
insufficient
sample size.
Bliss Independence
Bliss
Independence is
inaccurate
because it is prone
to false-positive
results and
assumes that
drugs work
independently
butthe way drugs
work in treatment
is not known
3
4. Purpose
To develop a novel
computational
algorithm to predict
synergistic cancer drug
combinations
4
6. Source Data: DREAM Challenge Data
◎Source for known synergy scores
○Largest and most comprehensive data set produced to
date measuring drug synergy
○118 drugs
○85 cell lines
○6,903 drug combinations
○586,705 drug pair on cell line combinations
◎Source for control cell line gene expression
○83 cell lines
○17,000 genes.
6
7. Source Data: LINCS L1000 Data
◎Source for drug gene expression data
○1769 drugs
○41 cell lines
○978 genes were measured
○22,628 genes are inferred from these 978
LINCS data is the largest dataset produced
measuring drug gene expression response on
a variety of cell lines.
7
8. Algorithm 1: Regularized Bilinear Regression
◎Most used to seeing linear regression, but needed a higher
dimensionality training algorithm as I was predicting synergy scores
of a pair of drugs on a cell line
◎Cell line gene expression data from DREAM Challenge
◎Drug on cell line gene expression data from LINCS L1000 database
◎Very large dimensionality dataset
○20 principal components for each drug & cell line
○400 principal components for each drug pair
◎Tensor product for D & C Matrix
◎Objective: to learn W matrix and predict Y (synergy score)
8
9. Algorithm 2: Up-Down Gene Analysis Regression
◎Regularized Bilinear Regression did not perform very well, thus I wanted to
improve on it by looking into the biological context of the data
◎For each drug experimented on each of 41 cell lines in L1000, categorized
each effect on gene as up-regulated or down-regulated
◎One 41 cell line X 911 gene matrix for each drug
◎Take mean over each cell line to find each gene’s probability of being up-
or down-regulated
◎111 drug up-down matrices total
◎For drug pairs add the values
◎Cell Line matrix kept the same as it was the control
◎Objective: learn W matrix and predict the same synergy scores
◎Performed Bilinear Regression again
9
10. Algorithm 3: Neighborhood Predictor
𝑆 𝐶𝐿,𝐷1,𝐷2 = 𝑆 + 𝑆 𝐶𝐿 + 𝑆 𝐷𝑃
𝑆 =
1
𝑁
𝑖,𝑗 ∈ 𝑆 𝑖,𝑗
𝑆𝑖,𝑗
𝑆 𝐶𝐿 =
𝑗 =1
𝑘
𝑑𝑗,𝐶𝐿 𝑆𝑗,𝐷𝑃
𝑗 =1
𝑘
𝑑𝑗,𝐶𝐿 𝑆𝑗
𝑆 𝐷𝑃 =
𝑗 =1
𝑘
𝑑𝑗,𝐷𝑃 𝑆𝑗,𝐷𝑃
𝑗 =1
𝑘
𝑑𝑗,𝐷𝑃 𝑆𝑗
𝑆 𝐶𝐿 =
1
𝑀
𝐶𝐿,𝑗 ∈ 𝑆 𝑡𝑟𝑎𝑖𝑛
𝑆 𝐶𝐿,𝑗
𝑆 𝐶𝐿 =
1
𝑃
𝑖,𝐷𝑃 ∈ 𝑆 𝑡𝑟𝑎𝑖𝑛
𝑆𝑖,𝐷𝑃
10
• Wanted to further improve on the results of
the Up-Down Gene Analysis Regression
• This algorithm is completely different than
last two—there is no regression
• Start by training a baseline predictor 𝑆, a
mean of all synergy scores in the dataset
• 𝑆 𝐶𝐿, cell line neighborhood adjustment
adjusted synergy score to the cell line by
looking at the synergy scores for all the
most cell lines to cell line in question
(similarity derived by correlation of gene
expression)
• 𝑆 𝐷𝑃, drug pair neighborhood adjustment
adjusted synergy score to the drug pair by
looking at the synergy scores for all the
most drug pairs to drug pairs in question
(similarity derived by correlation of gene
expression)
11. Performance Metrics
◎A test set was used to score algorithm
accuracy
○Leaderboard test set
○600 synergy scores
○167 drug pairs
○85 cell lines.
○400 scores could be used
◎Each algorithm predicted the synergy score
on these 400 drug pair on cell line
combinations
11
12. Results
12
Performance metrics for all three algorithms. RMSE is root mean squared error standard for many
statistical and learning problems. However, in this context it is not very important as it is not the
exact that are important, but the shape/patterns the real synergy scores take. Primary Metric and
Tiebreak metrics were obtained from the AstraZeneca Drug Combination Prediction DREAM
Challenge. Both measure the correlations of synergy scores (predicted vs. actual) within each cell
line. Correlation measures whole dataset Spearman correlation as a whole. Remarkably, all four
metrics improved as models got more complex and biological in nature.
13. Results
13
Graph of performance
of regularized bilinear
regression. X-axis is the
predicted synergy score
from the algorithm; Y-
axis is the actual synergy
score from the
leaderboard data from
the challenge. Line
through the middle
represents y = x, the
desired performance of
the algorithm. Lots of
clumping in the middle
shows that the model
was not the most
accurate but does show
promise that gene
expression can be used
to predict synergy
scores.
14. Results
14
Graph of performance of
up-down gene analysis
regularized bilinear
regression. X-axis is the
predicted synergy score
from the algorithm; Y-axis is
the actual synergy score
from the leaderboard data
from the challenge. Line
through the middle
represents y = x, the desired
performance of the
algorithm. Clumping in the
middle shows relatively
poor performance. Only
very few data points were
used as the scope of the
training data limited the
choice of ground-truth
variables.
15. Results
15
Graph of performance of
neighborhood predictor
X-axis is the predicted synergy
score from the algorithm; Y-axis
is the actual synergy score from
the leaderboard data from the
challenge. Line through the
middle represents y = x, the
desired performance of the
algorithm. Marked difference
between performance of this
algorithm versus other two
algorithms. Most data points of
all were used as feature data
did not limit choice of ground-
truth variables. The points
follow the curve of the y = x line.
16. Results
16
Similarity Grouping for Cell Line 22RV1
Based on cell line similarity matrix used in
Neighborhood Predictor. Shows that the
algorithm is not just attempting to guess a
relatively random number but also takes
into consideration the biological context.
The correlation matrix gets the similarity of
each cell line to all other 82 cell lines using
gene expression values to compare.
Similarity metric was able to pick up that cell
lines that are most similar to each other are
in the same disease area. Points are colored
according to disease area and several colors,
particularly red and green are clustering
together.
17. Discovery
◎Gene expression is a key component to predicting drug
synergy
◎Scientists should dedicate more time to studying gene
expression as it relates to synergy
◎Pharmaceutical companies can work to produce more gene
expression data on cell lines and drugs perturbations
◎Applying more of a biological approach improves
performance accuracy
◎Most other submissions to the DREAM Challenge only scored
in 0.1-0.15 range.
◎Most other submissions relied on a pure mathematical
approach, failing to take into consideration the biological
context.
17
18. Limitations
◎ Cell line control gene expression data
○ Only included 83 cell lines across varying tissues of origin
○ Next step is to use more diverse datasets with more cell lines
gene expression values from more tissues of origin
◎ Drug gene expression data
○ L1000 data only tested drugs on 41 cell lines
○ Only 6 cell lines overlapped with DREAM
○ Was unable to take drug on cell line gene expression
○ Error inherent in inferring other drug on cell line gene expression
values based on only six cell lines
○ If the same cell lines were used in both DREAM and L1000, could
have had a more direct comparison of gene expression drug and
cell line data to synergy scores
18
19. Further Research
◎Next step is to interpret these synergy
scores
○What is the cut off for a synergistic combination?
○What gene expression fold changes lead to higher
synergy scores?
◎Interpreting the weight matrix
○Gene expression fold changes in which genes tend to
lead to higher synergy scores
◎Important to test these combinations in
vitro to verify synergy
○Bridge between computational and experimental
19
20. References
Ali, S., Tonekaboni, M., Ghoraie, L. S., Satya, V., Manem, K., & Haibe-kains, B. (2017). OUP accepted manuscript.
Cerebral Cortex, (July), 1–14. https://doi.org/10.1093/cercor/bhw393
Bansal, M., Yang, J., Karan, C., Menden, M. P., Costello, J. C., Tang, H., … Shen, Y. (2014). A community computational
challenge to predict the activity of pairs of compounds. Nature Biotechnology, 32(12), 1–
12. https://doi.org/10.1038/nbt.3052
Costello, J. C., Heiser, L. M., Georgii, E., Gönen, M., Menden, M. P., Wang, N. J., … Stolovitzky, G. (2014). A community
effort to assess and improve drug sensitivity prediction algorithms. Nature Biotechnology, 32(12), 20–
23. https://doi.org/10.1038/nbt.2877
Dry, J. R., Yang, M., & Saez-Rodriguez, J. (2016). Looking beyond the cancer cell for effective drug combinations.
Genome Medicine, 8(1). https://doi.org/10.1186/s13073-016-0379-8
Foucquier, J., & Guedj, M. (2015). Analysis of drug combinations: current methodological landscape. Pharmacology
Research & Perspectives, 3(3), e00149. https://doi.org/10.1002/prp2.149
Gayvert, K. M., Aly, O., Platt, J., Bosenberg, M. W., Stern, D. F., & Elemento, O. (2017). A Computational Approach for
Identifying Synergistic Drug Combinations. PLOS Computational Biology, 13(1),
e1005308. https://doi.org/10.1371/journal.pcbi.1005308
Huang, H., Zhang, P., Qu, X. A., Sanseau, P., & Yang, L. (2014). Systematic prediction of drug combinations based on
clinical side-effects. Scientific Reports, 4(Figure 2), 7160. https://doi.org/10.1038/srep07160
Yin, N., Ma, W., Pei, J., Ouyang, Q., Tang, C., & Lai, L. (2014). Synergistic and antagonistic drug combinations depend
on network topology. PLoS ONE, 9(4). https://doi.org/10.1371/journal.pone.0093960
20
21. Acknowledgements
I would like to thank Dr. Christina Leslie, my mentor and principal
investigator of the Leslie Computational Biology Lab at Memorial
Sloan Kettering Cancer Center for allowing me to work in her lab as a
high school student and for providing me with so much advice and
guidance through this whole process. I would also like to Dr. Hatice
Osmanbeyoglu for always being available to talk about ideas or to
answer my questions and for training me in the beginning on
machine learning and cancer in general so I would be prepared to
work on this project. In general, I would like to thank the Leslie Lab
for welcoming me in as a high school student and for always being
available if I had questions. Lastly, I would like to thank my parents
who have supported me through all the ups and downs of this project
and always encouraged me to keep going.
21