Biomarker Discovery

Disease biomarker is a molecule that indicates changes in the physiology of a cell under diseased state and hence can be used as a diagnostic tool, therapy guidance and prognosis monitoring of diseases. For most common cell types and corresponding tumors there are no available antibodies in clinical diagnostics. The demand for new biomarkers is apparent in most areas of pathology and particularly within the field of cancer diagnostics. It is believed that the combination of liquid biomarker biopsy and AI technology will inevitably improve the accuracy of biomarker discovery.

Proteomics Approaches

Traditional Approaches.

Proteomics play an important role in drug development, relying on completion and annotation of the human genome and new refinements in the techniques to study proteins on the large scale, which has been applied to gain a better understanding of disease pathogenesis, to discover new and reliable biomarkers for early detection of diseases and to accelerate drug development. Traditional techniques include two dimensional polyacrylamide gel electrophoresis

(2-DE), mass spectrometry (MS), protein chip technology, phage display, activity based assays, two hybrid assays, isotope coded affinity tagging (ICAT) and multidimensional protein identification technique (MudPIT), and liquid chromatography/mass spectrometry (LC/MS).
In silico Approaches
Development of an in silico framework that is based on re-utilization of pre-existing epidemiologic and genetic data and that is aimed to identify candidate biomarkers indicative of diseases.

Data collection (publicly available database). Human Protein Atlas (HPA) program (www.proteinatlas.org) can be used to view protein profiles in a “gene/protein-specific” manner.

Figure 1 Data collection workflow

Create annotation database. Annotation is performed after immunohistochemical staining, defining cell populations present in the different tissues. Annotation parameters include intensity of immunoreactivity, fraction of positive cells, subcellular localization of the staining, and a free text box allowing for comments on the particular staining pattern.

Open source architecture (Java, PHP, and MySQL). The protein expression levels in all the included normal and cancerous tissues would be transformed into color codes for each cell type, which are finally assembled in a new database optimized for queries on expression patterns.

Discovery new biomarker. Enter multiple criteria to find candidate proteins with a high expression level in one tissue type, but low or negative expression level in another tissue type. Provide a systematic analysis of the millions of images included in the database and to generate lists of potential protein biomarkers.

Data Bases

Data extract techniques include lexical (pattern-matching) and linguistic (part-of-speech identification) for unstructured data sources.

Diseases Database.
Gene Ontology.
Genetic Association Database.
KEGG
NCBI Gene Expression Omnibus .
OMIM
PubMed
Medline

Data-driven Biomarker Discovery

In silico Literature Mining

Vast information is embedded in publicly available literature sources and other information databases relevant to specific diseases. Comprehensive analysis of these information has been conducted by advanced technologies.

Assertional Data .
Generation of massive volumes of highly accurate semantically consistent observational facts in the biomedical literature and other sources. Develop pre-curated vocabularies to enable lexical matching and to deal with the synonym variations across the data sources.
Intelligence Network (IN)
A semantically consistent form by applying comprehensive vocabularies and lexical matching approaches to yield a navigable database. All concepts are defined as a Concept Type. Each Concept was associated with a comprehensive set of synonymous terms.
Criteria Search
Identify important features for an ideal biomarker by setting the criteria.
In silico Screening for Biomarkers
MRI analysis, statistical analysis, and immunoblotting assays are applied to validate candidate proteins identified in the in silico study.

Figure 2 Network map used to guide the assertion generation for an Intelligence Network

Based on numerous data, intelligent tools have been developed implementing bioinformatics and machine learning methods for drug research and discovery. CD ComputaBio offers biomarker discovery and targeted proteomics services to researchers that want to benefit from our technology.