iNDX.Ai’s artificial intelligence (AI)-, machine learning (ML)-, and natural language processing (NLP)-powered applications iCorrelation Engine (iCE) and iQuery are highly intuitive software solutions for conducting complex analysis between multiomics and clinical data. The user- friendly interface of these applications obviate the need for advanced bioinformatics knowledge. More importantly, iCE and iQuery automate many aspects of data analysis – saving valuable time and allowing researchers to invest their energies in designing hypothesis and conducting experiments rather than spending countless hours in conducting mundane statistical tasks.
What is Multiomics
Multiomics encompasses analysis of the entire complement of biological information from various “omes”. Based on the scale of measurement, these “omes” may represent an entire complement of information for cell, tissue, or an organism and include
- Genome – the complete set of genes or genetic material
- Proteome – the entire complement of proteins that is or can be expressed
- Metabolome – the entire complement of metabolites produced by the source
- Transcriptome – depending on the analysis it is defined as the sum of all mRNA or total RNA
- Epigenome – the entire complement of chemical modifications occurring on DNA (e.g.methylation)
- Microbiome – the entire complement and diversity of microorganisms harboring any tissue or organism.
This can help researchers in deriving useful trends that can help better diagnoses, prevention, and treatment of different diseases. Multiomics can be instrumental in many research endeavors since it gives a comprehensive view of the disease and the affected patient in contrast to traditional research strategies that tend to study singular targets in isolation.
Why is Multiomics important for cancer
Multiomics has been instrumental in cancer research. Cancer is a highly heterogeneous disease and therefore requires comprehensive tools that can dissect the intricate complexities that drive this disease. Multiomics is a great tool for dissecting cancer heterogeneity – a phenomenon that makes it incredibly difficult to treat this disease. This heterogeneity exists at multiple levels. The first level of heterogeneity occurs in cancer itself where its origin (eg. breast vs. lung, DCIS vs. LCIS etc.), molecular drivers (p53 vs. PTEN, ER + vs. ER -vs. TNBC etc.), inherent aggressiveness, and surrounding environment (blood supply, tumor microenvironment it grows in etc.) dictate its behavior. The second level of heterogeneity is at the level of patients where their genetics and environment can determine their susceptibility to the disease and their response to the treatment. In any case, this heterogeneity arises from differences in different omic components such as genome, proteome, metabolome, transcriptome, epigenome, and microbiome. The complex interactions between various omic components are often missed when they are studied in isolation.
The appreciation of these interactions and the success of multiomics in dissecting them has led to widespread recognition among scientific and clinical communities regarding the customization of treatment based on the individual “omic” characteristics of cancer and the patients. The use of multiomics has made it easier to stratify patient subpopulations according to biomarkers and achieve higher precision in diagnosis, prevention, and treatment of cancer. The use of multiomics in biomarker discovery facilitating treatment decisions is now quite common. For instance, multiomics strategies are frequently employed in biomarker discovery and gauge patient response for immunotherapies (Pembrolizumab response MSI-high vs. MSI-low), chemotherapies (PARP inhibitors based on BRCA status), and targeted therapies (presence of constitutive tyrosine kinase mutations like (constitutive RTK like EGFRv III and BCR-ABL)
What are the challenges in the analysis of multiomics data
The biggest challenge in Multiomics is data analysis. One of the worst nightmares for any biologist is to conduct expensive and highly complex high throughput experiments and then have a laundry list of targets all of which seem to equally attractive! How does one chose which ones are relevant? An RNA-Seq experiment alone, for instance, can result in several hits which may seem relevant at first. But then further analysis with corresponding proteomics data may reveal a different story altogether. When information from other omic components and clinical sources is superimposed on the analysis, newer trends may become apparent which are not identifiable when only one of these datasets is analyzed. But how does one carry out such correlative analysis?
We at iNDX.Ai recognize challenges that are faced by biologists and have developed two user-friendly applications – Integrated Correlation Engine (iCE) and iQuery. The goal of these applications is to make life easy for scientists working on complex oncology projects and accelerate the rate of scientific discovery. These two tools have been designed to enhance the capabilities of investigators in conducting a cross-functional analysis on omics and clinical data. The AI-, ML-, and NLP-powered technologies running these applications make for a highly intuitive software solution, obviating the need for advanced bioinformatics knowledge. More importantly, iCE and iQuery automates many aspects of data analysis – saving valuable time and allowing researchers to invest their energies in designing hypothesis and conducting experiments rather than spending countless hours in conducting mundane statistical tasks. Because these tools are cloud-integrated, it gives them immense power and flexibility in conducting complex statistical and correlative analysis between multiomics and clinical data.
What is Integrated Correlation Engine (iCE) and how can it help you
Integrated Correlation engine (iCE) is a bioinformatics application that uses novel data correlation models that facilitate the investigation of relationships between multi-omics and clinical outcomes datasets. This results in an improved ability to associate biomarkers with clinical phenotypes of interest. By executing an artificial intelligence (AI)-powered integration, organization, visualization, and analytics of multi-dimensional data; the Integrated Correlation Engine (iCE) automate the biomarker discovery process. Graphical outputs greatly assist in the interpretation of complex analyses done on multiomics and clinical data. These features provide valuable biological insights and mine for patient-specific biomarkers in an unbiased manner as a function of treatment duration and dose.
We describe here some useful features of the iCE platform through some real-world examples:
a) iCE dashboard can display a relative expression of ten different genes in three cohorts of patients (progressive disease, partial response, and stable disease). The current output shows information at a specific time point derived from an RNA-Seq experiment. Notice drop-down menus and interactive scale for adjusting a number of markers and types of analysis (modality). This enables mining of similar or new gene expression changes from other types of experimental queries. Analysis like these form a basis for quick identification of biomarker trends for a particular patient cohort (e.g. responders vs. non-responders)
b) iCE allows the user to normalize the data as per their requirements. Also recommends normalization parameters that will give the best results.
c) iCE is powered to conduct multi-omics correlation analysis enabling researchers and scientist to identify biomarkers and stratify patients according to their response and phenotype
d) iCE can conduct complex analysis and displaying data in several graphical outputs. Shown here are two representative graphical output format. On the left notice, a volcano plot showing a correlation between p-value and the change in gene expression measured in fold change in for a particular experiment. Shown on the right is a dot-plot distribution of samples (cohort of patients with observed metastasis)
e) iCE is powered to conduct principal component analysis (PCA) on multi-omics and clinical data – enabling researchers to identify powerful clusters of genes (notice marker cluster and PID cluster) associated with a particular phenotype.
What is iQuery and how can it help you
iQuery is an advanced data interrogation engine based on natural language processing (NLP) and free text formatting techniques. In other words, this allows clinical investigators and scientists to run detailed queries across clinical, biomarker and imaging datasets in plain English. The simple user interfaces and rule-based drop-down menus allow the biologist to design the queries without any prior programming skills. iQuery can empower any researchers to ask a plain text question regarding a biological relationship without having to worry about generating programming constraints to accomplish this task.
For instance, an attractive feature of this platform is called Query Builder and allows you to frame a complex biological question by simply specifying information into five basic parameters or rules. In this case, the question is defined by selecting information from five columns – category, attributes, operation, values, and conditions. These logic rules can be saved for doing a similar analysis of different datasets.
In Summary, these tools are a valuable resource for conducting complex bioinformatic analysis and deconvolute multiomics data. The user-friendly dashboards and intuitive user interface are powerful features that allow the researchers to efficiently conduct complex bioinformatics analysis.
Based in Silicon Valley, we are a highly motivated team of entrepreneurs, doctors, engineers, and data scientists with a mission to develop innovative software products that can help our collaborators find a cure for cancer and make a global societal impact.
If you would like to learn more about how we can help you overcome your clinical trial hurdles please feel free to contact us here.