We are committed to developing global original life science technology.

  1. Home
  2. >
  3. Technology

AI-based Digital Pathology

MBMbio digital pathology platforms quantitatively analyse the chromatin organisation and DNA ploidy in tumor nuclei, and tumor-stroma fraction. Our technology is powered by artificial intelligence(AI). We use machine learning and deep learning algorithms to discover new biomarkers in whole slide images. These biomarkers have been clinically verified as pan-cancer prognostic biomarkers for colorectal, prostate, gynecological, lung cancers and other solid tumors.

Nucleotyping (chromatin heterogeneity)

Chromatin organisation is quantified by computing the entropy of pixel grey levels in a subregion of a nucleus, which reflects epigenetic modification of DNA and ultimately gene expression. Nucleotyping results in a biomarker measurement termed as chromatin value that is used to classified tumors as homogeneous or heterogeneous.

Ploidy (DNA content)

Ploidy reflects chromosome copy number in tumor cells. Abnormalities of cellular DNA content (polyploidy and aneuploidy) have long been associated with tumorigenesis and recently associated with immune response modulation. 

Stroma (Tumour-stroma fraction)

Stroma measures degree of stroma invasion surrounding cancer cells. Tumor stroma is comprised of cancer-associated fibroblast, endothelial cell, myloid cell, pericyte, lymphocyte and extra-cellular matrix. It plays an important role in tumorigenesis, tumor invasion and metastasis.

*Nucleotyping/Ploidy/Stroma were licensed from Room4 (UK), and originally developed by the Institute of Cancer Genomics and Informatics (Norway).

Red Blood Cell-Based Liquid Biopsy

One of the major advances in the field of precision cancer treatment is the advent of non-invasive liquid biopsy. Most existing technologies involve the detection of circulating tumor DNA (ctDNA) and circulating tumor cells (CTCs) from blood plasma. However, a major bottleneck with this approach is that these biomarkers are shed from tumors only at a later stage of the disease. This means that we will miss the chance to detect the disease at its earliest stages, when the chance of effective treatment is the highest.

At MBMbio, we have developed a Red Blood Cell(RBC)-based platform, a new approach to liquid biopsy that addresses the inherent problems associated with methods based on ctDNA and CTCs. We are currently performing clinical validation studies on the RBC platform in stage I non-small cell lung cancer patients.


Sequencing 2’-O-Methylation of  RNA

RNA epitranscriptomics, the study of RNA modifications at scale, is the new frontier of understanding biological system functions.

Our scientists have recently developed an accurate and efficient method for detecting and studying 2’-O-Methylation from a transcriptome-wide point of view. NJU-Seq (Nm-site-Judge-Universally Sequencing) is the first enzyme-based high-throughput 2’-O-Methylation site screening method with single base resolution.

The NJU-Seq method relied on helicase and hydrolase activities of the RNase from Mycoplasma genitalium (MgR) on ssRNA from 3′ to 5′ direction, which produced the digested products at every Nm+1 site. Using this tool, we have identified thousands of new 2’-O-methylation sides in mice, human and even viral RNA molecules.  

Schematic representation of NJU-Seq protocol.

Sequencing Method

Sequencing Principle

Described Applications

Single step digestion with MgR, a purified Nm site specific hydrolase
rRNA, mRNA, snoRNA, viral mRNA
Multi-step sodium periodate oxidation and sample purificiaion
Based on resistance of RNA methylation sites to random alkaline cleavage
Abortive RNA reverse transcription at low dNTP concentration

Comparison of NJU-Seq to other 2’-O-Methylation sequencing methodologies