Page 372 - Special Topic Session (STS) - Volume 4

STS1080 Mahmoud Rafea et al.
biological systems understanding regarding the domain of health and disease.
Biomarkers are considered a great director in the development of new
treatment strategies [30].
One of the most important applications of specific biomarkers is to find
the tumor at an early stage even before clinical symptoms are developed.
Patients with cancerous disease can be treated efficiently when detected in
early stages [31]. This would certainly increase overall survival. The World
Health Organization (WHO) proposed that “millions of cancer patients could
be saved from premature death if early detection and treatment were
available” [32].
In addition to early diagnosis, evidence-based medicine can profit from
biomarkers knowledge providing; selection of the optimal therapy and
improving prognosis of diseases. The greatest potential for enabling
biomarkers for cancer lies in improving the technology for protein biomarker
discovery. Protein biomarkers should be found in a minimally invasive liquid
biopsy such as a simple blood sample. Nevertheless, does blood contain
enough information? Numerous researchers in this era work to find protein
cancer biomarkers of clinical utility [33, 34, 35, 36].
In [37] they conclude that the antigens exist in the RBC cytoplasm have
inversely proportional to immune tolerance. When the antigens in RBC
increases the antibodies in plasma decreases and vice versa, where Ag
represents the antigens in RBC, and Ab represents antibodies in plasma. Also,
they conclude that RBC has a dynamic store of: body antigens (Tissue Specific
Antigens (TSA)), food antigens, environment antigens, bacterial commensal
antigens, and disease antigens whether microbial, viral, or tumors. This store
is known as: Erythrocytes Dynamic Antigens Store (EDAS).

4. Applying Machine learning algorithms on EDAS
In order to make maximum benefit from this discovery, computer
processing capability and computer knowledge processing capability can help
to profit this discovery. To this endeavor, a random generation of EDAS was
described in [38]. Meanwhile, the generation of EDAS model was very simple
and did not reflect the real EDAS. It was based on classifying proteins into
normal and abnormal, only, without specifying the nature of these proteins.
Also, in [37, 38] they proposed a technique to discover biomarkers of diseases
based on EDAS. However, these algorithms are very simple, and do not match
with reality EDAS. Also, they did not show which disease or set of diseases can
be applied on? They used one category of diseases. And, they do not make
any experiment to verify the model.
Recently, in [4], researchers developed the random generation of EDAS.
This developing is based on proposing a new mathematical model to abstract
the problem computationally with richer knowledge. This new random
generation of EDAS data simulate reality. The generated EDAS data consists

361 | I S I W S C 2 0 1 9