See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/350974126

The Cell Cycle Model: A Comprehensive Review and Extension Based on

Machine Learning

Article in International Journal of Education and Management Engineering · April 2021

DOI: 10.5815/ijeme.2021.02.02

CITATIONS READS

0 114

1 author:

Mustafa Kamal Pasha

Lincoln University
7 PUBLICATIONS 12 CITATIONS

SEE PROFILE

All content following this page was uploaded by Mustafa Kamal Pasha on 12 October 2022.

The user has requested enhancement of the downloaded file.

I. J. Education and Management Engineering, 2021, 2, 13-24
Published Online April 2021 in MECS (http://www.mecs-press.org/)
DOI: 10.5815/ijeme.2021.02.02

The Cell Cycle Model: A Comprehensive Review

and Extension Based on Machine Learning
Mustafa Kamal Pashaa, Khurram Munawara, Asma Talib Qureshib
a
Department of Environment, Society and Design, Lincoln University - New Zealand
b
Department of Healthcare Biotechnology, National University of Sciences and Technology – Pakistan
E-mail: kpasha2003@gmail.com

Received: 24 November 2020; Accepted: 14 January 2021; Published: 08 April 2021

Abstract: The cell cycle is a conserved process comprising of an organized series of interdependent and cross
regulatory events that lead to controlled cell growth and proliferation. Genomic and volume regulatory processes are of
special interest as they decide the fate of cell cycle. Signaling cascades including MAPK, PI3K, Sonic Hedgehog, Wnt
and NOTCH signaling pathways are few well known conventional players contributing in controlling the cell cycle
progression through different phases by expressing certain proteins. Moreover, the unconventional volume regulatory
players exert influence by regulating membrane potential that is determined by ions influx or efflux across the plasma
membrane via ion channels, controlling water movement and ultimately contributing to volume increase in growth
phases of the cell cycle. Both of these players are interlinked, therefore, in order to establish a better understanding of
the interdependence of these players, principles of machine learning were applied on data obtained on cell cycle. The
data was processed by using neural networks and it shows that a significant understanding of conventional regulators is
available in the literature and it has been under the limelight as well. However, when it comes to unconventional
volume regulatory players, a limited understanding is available. Moreover, the precise role of each component and its
interdependence with other is not yet fully understood. Due to which, they are not clearly evaluated for their potential
role as cell cycle control elements for therapeutic purposes. Therefore, this study aims to summarize the data on cell
cycle that is obtained through machine learning and to discuss the advances in cell cycle modelling mechanisms and
designs that are based on different mathematical algorithms. Thus, this review will provide a basis to clearly understand
and interlink the discoveries on cell cycle so that a comprehensive cell cycle model could be built which, if manipulated
can be used for therapeutic purposes by identifying the least explored regulatory control elements.

Index Terms: Cell cycle, Intelligent modelling, computational modelling, and role of Ca2+ signaling, Artificial Neural
Network, machine learning

1. Introduction

The cell is a fundamental unit of life and plays a crucial role in organ and system development, transportation and
storage of biomolecules, gene expression, signal transduction, and empowerment of molecular machineries [1]. The
process of a cell dividing into two daughter cells is known as the cell cycle, which is a universal and complex process
and is tightly regulated by different regulatory proteins that either allow or limit its progression [2]. A cell progresses to
cell division by receiving growth and proliferative signals from the extracellular environment. In response to these cues,
a cell quits G0 phase and enters into the G1 phase of the cell cycle. The G1 is one of the two growth phases in the cell
cycle where cell increases in size and accumulates within itself sufficient nutrients to provide energy during a cell cycle.
This volume growth is an important regulatory step as it decides the progression of cell into subsequent phase. When all
the required conditions are met, a cell progresses to S phase that is the synthesis phase of DNA. Following the S phase
is the second growth phase called G2 phase where cell replenishes its energy reserves and grows in size so that it could
enter into the mitotic phase, which terminates this cycle at cell division by passing through four substages, starting from
prophase and ending at telophase [3]. The first 3 phases G1, S and G2 comprise the longest period in cell cycle known
as interphase [4]. There are few mammalian cells, for instance epithelial cells, which continually grow and divide while
other cells stay in quiescent phase and perform normal metabolic functions like muscle cells or neurons. Investigators
have postulated and later proved that these cells stay longer in G0 phase and upon receiving stimulus they continue
proliferation or differentiation. This longer stay in the quiescent phase has been distinguished as G0 phase distinct from
G1 phase [5].
Upon receiving mitogenic signals, changes in cellular dynamics are observed due to activation of certain signalling
pathways which ultimately cause transcription of proliferative genes by expressing transcription factors like FOS, JUN,

Copyright © 2021 MECS I.J. Education and Management Engineering, 2021, 2, 13-24
14 The Cell Cycle Model: A Comprehensive Review and Extension Based on Machine Learning

cMYC etc. Some of these signalling cascades are MAPK pathway, PI3K/Akt/mTOR pathway, NOTCH pathway, SHH
and Wnt signalling pathway. All of these pathways play a significant role in cell cycle progression by directly
influencing the genetic machinery and transcribing the genes and expressing proteins essential for cell progression
through different cell cycle phases. The most crucial regulatory proteins responsible for transition and transversion of
cell cycle checkpoints are Cyclins and Cyclin Dependent Kinases (CDKs) along with their inhibitors, and tumor
suppressor genes p53 and pRb and associated regulators [2]. These regulatory signalling molecules are considered to be
the conventional drivers in cell cycle control throughout its four basic stages to complete the cycle of division, namely:
G1, S, G2 and M phase [6].
As research on the cell cycle progressed over years, new regulatory control elements were identified. Initially, the
cell cycle was understood to be dependent on a number of different phases that a cell goes through during cell division;
however, later, other vital aspects were found. One of these was the role of bioelectricity in driving the process of cell
division. This bioelectricity was observed due to influx and efflux of different ions through ion channels that are present
on the cell membrane as well as on the nuclear membrane [7]. This flow of ions across the plasma membrane
establishes a membrane potential which has been linked with the regulation of the cell cycle since long and are also
involved in cancer development and progression. However, how they control cell cycle has not yet been fully elucidated
[7].
All these pathways and regulatory elements are intricately interlinked as they all are performing the same function
i.e. driving the cell to division. But no study has been conducted till date that would have incorporated these details in a
single model to provide a better understanding of the cell cycle. Therefore, in order to get a bigger picture, Artificial
Intelligence should be introduced in the domain of biological sciences. Hence, this study aims to evaluate the modelling
of the cell cycle of a living organism through machine learning. Previously, the new technologies were only used in the
areas of industry and education. But the recent years have seen an increase in the need of integrating the modern age
equipment in the field of health and biological sciences [8]. Owing to the increased use of computers and internet
sources, machine learning is the idea of vast volumes of data being produced every day in various fields [9,10].
Therefore, approach of machine learning is utilized in this study to evaluate and model the cell cycle. Other than
computer, the devices like antennas and wireless devices can also be used for the collection of data. As the data
collected through these devices would have large benefits, including the domains of personal wellbeing, and biological
sciences [11-16]. Therefore, this study utilizes the effectiveness of Neural Networks in order to better understand this
whole process of cell division. These Neural Networks work on the similar principle as that of human brain by
processing the information in different layers. These layers are input layer, hidden layer and output layer. Number of
layers in hidden layer can be increased depending on availability and complexity of the data. As this study evaluates a
huge data set on cell cycle and its regulatory control elements, therefore, we have utilized the effectivity of machine
learning and artificial intelligence so that all the data could be processed in the best possible manner through neural
networks. The findings of this study would enable researchers to critically evaluate the influence of their research
outcomes by incorporating them in a comprehensive cell cycle model. This would on one hand help them in validating
their current findings and on the other hand will help in identifying new cellular targets.

2. Methodology

The methodology adopted for this research is shown in Fig. 1.

Fig. 1 Methodology based on Machine learning adopted for cell cycle review and modelling

2. 1 Data Retrieval Through Machine Learning

Data required for conducting this research was collected through an extensive literature retrieval from different
search engines including Web of science, ScienceDirect, and Scopus, etc., through a combination of index terms.
Research from high index papers published in top-tier journals was selected and was further processed by using
machine learning and artificial intelligence. A number of keywords were used to obtain relevant literature on cell cycle.
The keywords used for this research includes “Cell cycle, Intelligent modelling, computational modelling, role of Ca +2
signalling, Artificial Neural Network, cyclin and cyclin dependent kinases, cell cycle machine learning”.

Copyright © 2021 MECS I.J. Education and Management Engineering, 2021, 2, 13-24
 The Cell Cycle Model: A Comprehensive Review and Extension Based on Machine Learning 15

2. 2 Modelling and Analysis via Neural Networks

For the collection and review of the complex data on cell cycle, Multilayer Artificial Neural Networks were used
in four phases. These layers, as shown in Fig. 2, include, Data Extraction, Pre-processing and preparation, Modelling
and Evaluation. The neurons links were made in between these phases. The aim of these complex networks is to find
the weight of the data on research domains from cell cycle. Using this ANN model all the data is collected and
evaluated in this research.

Fig. 2 ANN Model for the data collection

AI based layers are also used for the data classification and evaluation. The layers, as shown in Fig. 3, includes,
convolution, pooling, and classifier. The aim of these complex networks is to find the weight of the data. This model
helps to collect detailed data on the cell cycle.

Fig. 3 AI based Model for the data Evaluation

After the data pooling through AI, the data was classified in different cell cycle layers based on ANN. These layers
define each stages of cell cycle separately. The framework is shown Fig. 4. Layers adopted in this framework includes
input, convolution, activation and full connection with the output layers,

Copyright © 2021 MECS I.J. Education and Management Engineering, 2021, 2, 13-24
16 The Cell Cycle Model: A Comprehensive Review and Extension Based on Machine Learning

Fig. 4 Data classification based on ANN

3. Results

3.1 Neural Networks

VOSviewer software generated a network from output data of ANN Model based on “Visualization of Similarity”
as given in Fig. 5. The nodes of this network represent the domains explored for cell cycle while their size represents
the weight of the data, or in other words the amount of work done under these domains. The interconnectedness of these
research areas is represented by lines between the nodes. The most repetitive keywords found in the articles include
cancer, ca2 signaling, inhibitor, differentiation, cell cycle, apoptosis, kinase, neuron, etc. This clearly shows that
research on cell cycle is mostly related to genomic machinery and it lacks the concept of comprehensiveness and
inclusion of volume regulatory control elements. Therefore, this study presents a review on the available literature on
unconventional elements of cell cycle control by focusing on volume regulatory components which is the main
exploratory domain of this research.

Fig. 5 Most repetitive keywords used in research articles related to cell cycle

Copyright © 2021 MECS I.J. Education and Management Engineering, 2021, 2, 13-24
 The Cell Cycle Model: A Comprehensive Review and Extension Based on Machine Learning 17

3.2 Data Derived through Machine Learning On cell cycle

Based on the results of ANN Model, data on cell cycle was retrieved through machine learning from online sources
and is discussed below
3.2.1 G1/S phase transition
According to data pool [9, 17, 18] collected through machine learning, cells stay in G 0 phase and perform several
metabolic functions unless they receive a signal from extracellular matrix for progression. This overall progress from G 0
to G1 is tightly controlled by a series of growth factors, including Platelets-Derived Growth Factors (PDGF), Fibroblast
Growth Factors (FGF), Insulin like Growth Factors (IGF) and insulin which determine the regulation of the cell cycle
phases from one to another. The data also revealed that if the cell does not encounter these growth factors, it then
remains in quiescent state [9]. Upon receiving growth factor signals, the cell enters G1 phase and performs a number of
distinct activities. It activates pathways which establish Ca +2 signalling in waves that plays a role in controlling the cell
cycle machinery, increment in cell volume to hold DNA and other copied cellular components, and the synthesis of
proteins required for entry into S phase as shown in Fig. 6. According to Pledger, this transitional state occurs late in G1
after satisfying the requirements of the volume control checkpoint and the absence halts the transition into S phase or its
deregulation can lead to abnormal cellular transformations and cancer conditions [17]. Moreover, p53 is found
responsible in the database for controlling this transition in the case of DNA damage [18] as the controller of cell cycle
arrest and DNA repair.

Fig. 6 Phases of Cell Cycle (Interphase and Mitotic phase) (Ravindra B. , 2006)

3.2.2 Establishment of Ca2+ signaling

Ca2+ is found as a ubiquitous molecule in the data pool. According to Rossow, 1979, it controls cell volume,
bioelectricity and cytoskeleton and servring in the process of coordinated cell growth in G1 and G2. It was found in
literature that the Ca2+ establishment is achieved through initial cell shrinkage and its influx relies on activation of
various tyrosine kinase receptors. ANN also revealed that receptor-ligand binding activates the PI3K and the
extracellular signal-regulated kinase1/2 (ERK ½), which in turn activates the potassium channels that have kept the
membrane potential in a depolarized state to eventually change it to a hyperpolarised state through the activation of
calcium channels in mid G1 and towards the transition of G1/S phase [19]. The study of Bartek 2001, states that during
mid G1, Phospholipase C is activated, which mediates the release of Ca2+ from the endoplasmic reticulum into the
cytoplasm that maintains and regulates the activity of the Ca2+ mediated K+ channels leading to cell volume shrinkage.
The hyperpolarized state is thus maintained, mediated by further entry of Ca2+ into the cytoplasm leading to the
activation of the transcription factors such as JUN, FOS and c-MYC, which transcribe CDKs and cyclins, such as the
p15, p16, p17, p18 and p19 thus inhibiting the activity of the CDK inhibitors (p21 and p27) [20]. Bartek 2001, also
states that once the Ca+2 is established, volume increase along with early immediate gene expression takes place.
3.2.3 Increment in Cell Volume and Its Control
The phenomenon of volume control has been speculated in the data sets of CNN, to be an important part of the cell
cycle. According to Massagué, 2004, nucleus to cytoplasm ratio is found as an important measure in monitoring the
increase in cell volume. A continuous increase is evident in the cell volume starting from the G1 phase all the way to
the M phase. According to collected data, the role of the cytoskeleton in the regulation of the cell cycle has not been
proven yet and there is a perplexed opinion regarding the regulatory mechanisms since the inhibition of the cytoskeleton

Copyright © 2021 MECS I.J. Education and Management Engineering, 2021, 2, 13-24
18 The Cell Cycle Model: A Comprehensive Review and Extension Based on Machine Learning

through drugs has shown no effect on the cell volume changes. The size of the nucleus has been said to be directly
regulated by the nuclear membrane in the data sets. The important catch data provided is that once the cell actively
begins the transcription phase, there is a need to control the volume of the cell and its components. According to
Massagué, 2004, this is achieved by the involvement of the chloride ions which actively regulate the cell volume. The
expression of the EAG2 (K+) channels have been found in data to be implicated in the control of the M phase by means
of regulating the expression of cyclin B1 through the p38 MAP Kinase, as shown in Fig. 7 [21].

Fig. 7 Alterations and Proliferations in Cell cycle (Source: Massagué, J. 2004).

The question is, which ions or proteins are responsible for maintaining ionic transport balance for the control of
cell cycle? It is proposed earlier in the articles, that Ca++ oscillations have a crucial role in the control as a “life and
death signal” [22]. Later, it was suggested in the studies, that the oscillations also follow and elicit the NKCC and NHE
channels for homeostatic perturbations for planned activities of cell intracellular signalling machinery. These
oscillations have been said to produce the membrane potential (Vm) changes as a by-product (as they are maintained
through the activity of a range of ion channels), which in itself is necessary for proliferation. Besides this, a variety of
K+ channels have been implicated in the regulation of proliferation and cell cycle progression. Mechanisms of the
activation of many of these channels are currently unknown [23].
As stated earlier in the data, before volume increase, cell proliferation may require transient cell shrinkage initially,
which is accomplished by the activation of Cl- and K+ channels. As the electrochemical equilibrium activity of Cl- ions
is above the threshold inside the cell, the activation of Cl- channels is responsible for Cl- exit and thus resulting in
depolarization. If Cl- exit is paralleled by active K+ channels, then there is net exit of KCl salt, responsible for cell
shrinkage. Conversely, the cell growth requires an increase in the K + concentration inside the cell. The inward rectifier
uptake channels of K+ activate simultaneously with mechanosensitive channels that are activated by the compression in
the plasma membrane that has undergone shrinkage, and help in bringing in higher concentrations of K + and water,
which increase the turgor pressure, which is needed for cell growth [24].
Data from studies confirmed that the Kv10.1 cause a reduction in the current on the cell membrane. The reduced
current is known to be associated with the mitosis-promoting factor (MPF- p24) and Na+. The K+ concentrations also
regulate the entry of Ca2+ inside the cell by ensuring that the membrane potential is negative enough to allow the entry
of Ca2+ [25]. According to the study of Lang 2007, the Kv1.3 channel in conjunction with the KCa3.1 (a Ca 2+ dependent
K+ channel) works towards the cell growth and proliferation. Ca 2+/CaM is required at two points during the re-entry
from quiescence, early after mitogenic stimulation and later near the G1/S boundary. Cell volume not only participates

Copyright © 2021 MECS I.J. Education and Management Engineering, 2021, 2, 13-24
 The Cell Cycle Model: A Comprehensive Review and Extension Based on Machine Learning 19

in the regulation of cell function by hormones, but also regulates hormone release. The release of several hormones is
triggered by cell swelling, conversely inhibited by cell shrinkage. The link between cell volume and hormone release is
ill-defined in the data sets but partially involves cell volume sensitive alterations of cytosolic Ca2+ activity [25].
3.2.4 G2/M phase transition
Studies revealed that, once DNA duplication is completed, the cell proceeds towards the G2 phase where it
prepares itself for division into two daughter cells in M phase. A series of events, as shown in Fig. 8, beginning with
prophase and then prometaphase, metaphase, anaphase and lastly cytokinesis takes place-which divides the cell into two
genetically identical copies during the M phase. A clear role, found in literature, of the dynamic distribution of Ca2+ and
Ca/CaM is seen in S/G2, G2/M and during M phase. Towards the end of M phase, it was found in studies that, Ca/CaM
concentrates itself below the membrane and helps cleave the cell into two. Just prior to M phase, DNA damage
checkpoint have the responsibility to give a “Go” signal for advancement to checks whether the DNA is intact or has
any mutations to repair. A study by Sanchez et al., stated that it would be catastrophic if cell proceeds to divide with the
damage. Studies also revealed that, during G2, mammalian Cyclin B/CDK2 complexes are held in an inactive state by
phosphorylation of CDK1 at the two negative regulatory sites, threonine 14 (Thr14) and tyrosine 15 (Tyr15) [26].

Fig. 8. Data collected through machine learning on G2/M Transition

The data suggested that a study by Kahi et al., 2003, the Ca 2+/CaM is implicated in the G2/M transition, M phase
progression, and exit from mitosis. During the M phase the MPF increases the selectivity and rectifies the current,
promoting the loss of K+ from the cell. Moreover, calcineurin was found as Ca 2+ waves regulators in G2/M transition.
Another prevalent enzyme, in addition to cdc25, CAMKII is found as important for G1 progression and G2/M transition.
As demonstrated in glioma cells, a drastic volume decrease (or distribution per se) occurs during the M phase to reach a
preferred volume state in the division. This relates to the chromatin and cytoskeleton condensation and
depolymerization in M phase [27].
3.2.5 Role of Ca2+ in cell cycle
There are a number of different classes of calcium receptors, found in literature through machine learning, which
play a crucial role in the maintenance of cellular homeostatic conditions and regulation of cell cycle by mediating the
entry and exit of the calcium ions and also stimulating the intracellular release of calcium. Even it was found that the
endoplasmic reticulum has calcium receptors, which help in storing calcium and releasing it as and when needed inside
the cell during the cell cycle. The detailed classes of calcium receptors were reported by the data which includes the
Voltage Gated Calcium Channel (VGCC) which is itself comprised of three different families of receptors, including
the Cav1, Cav2 and Cav3. The other classes of calcium receptors include the Receptor-Operated Calcium Channels
(ROCCs), the Ryanodine Receptor (RyR) present at the endoplasmic reticulum and the Inositol-1,4,5-trisphosphate
receptor (IP3R), as shown in Fig. 9 [28].

Data Pool
on Role of
Ca2+ in Cell
Cycle

Fig. 9 Role of Ca2+ in Cell Cycle (Source: Villalobos et al., 2001)

Copyright © 2021 MECS I.J. Education and Management Engineering, 2021, 2, 13-24
20 The Cell Cycle Model: A Comprehensive Review and Extension Based on Machine Learning

Data collected through machine learning showed that, the Ca2+ channels control and regulate the cell cycle. In
excitable cells such as the muscles, neurons and the endocrine cells, voltage gated calcium channels are used for the
entry of Ca2+. The increase in Ca2+ concentration inside the cell leads to the phosphorylation of the Mitogen-Activated
Protein Kinase (MAPK), ultimately leading to the progression of the cell cycle. The downstream processes following
the entry of Ca2+ ions are diverse and include expression of a number of different genes, depending on the cell type. The
concentration of Ca2+ in the resting stage of the cells is very low (around 10-7 M), while in the calcium stores such as
the endoplasmic reticulum and the extracellular matrix, Ca 2+ concentrations are much higher (around 10,000 times
higher, approximately 10-3 M). The maintenance of this gradient is ensured by the efflux of calcium from the
intracellular organelles [29].
It was also confirmed in articles [27-30] that, Ca2+ interacts with the Cyclin-Dependent Kinases (CDKs) to regulate
the cell cycle. The CDK family is considered to be important in the transition of the cell through the different phases of
the cell cycle and also in the maintenance of the different phases. As the data suggested, CDK4 and CDK6 are found as
important in the G1 phase, CDK2 is important in the G1 as well as the S phase, and also speculated to be a part of the M
phase, while CDK1 is predominantly reported to be important in the M phase. Ca 2+ in complex with Calmodulin (CaM)
interacts with the CDKs and controls their activity throughout the cell cycle. Ca 2+ and CaM together regulate the
expression of the CDK1, CDK2 and the Cyclin B in human cells (particularly reported in the T lymphocytes). Moreover,
it was found that the activation of the stored Ca2+ (SOCE: Store Operated Calcium Entry) activates CaM protein, which
in turn leads to the blockage in the activity of Cyclin A and E [30].
According to the study of Se et al., 2004, Ca 2+ oscillations also play a role in the gene expression. This has been
documented in the case of the early and late gene expressions in the G1 phase. In the early phase of G1, Ca2+ affects the
expression of the Serum response element (SRE), the Cyclic AMP Response Element (CRE), MYC, JUN and FOS
genes, which are all important for the proliferation of cells. The activity of Ca 2+/CaM leads to the activation of
CDK4/Cyclin D1 complex, which is involved in the regulation of the retinoblastoma protein (RB1), which is one of the
main inhibitors of the DNA synthesis process. The RB1 is found as responsible for interaction with T2F transcription
factor for the inhibition of cell cycle. However, the regulatory activities and the phosphorylation of the RB1 protein
leads to the transition of the cell cycle from the G1 to S phase. According to Se et al., 2004, this transition is mediated
by the regulatory activity of Cyclin D1 and the Ca2+/CaM pathway. Here, the RB1 is inhibited and the p21 and p27
(better known as CDKN1A and CDKN1B) are also negatively regulated [31].
Next, the role of Ca2+ has been reported in the transition from G2 to M phase. It has been seen that Ca 2+/CaM
regulate CAMKII in the G2 phase [59] and lead to the CAMKII mediated phosphorylation of the Microtubule
Associated Protein 2 (MAP2), which leads to the inhibition of the microtubule polymerization [32].

4. Advancements in Cell Cycle Modelling

Mathematical modelling and simulation of the cell cycle started a couple of decades ago. This includes, high
throughput screening, modelling and simulations, topological interactions for the prediction of the system function and
many others. None so far included the machine learning processes, including the training and test sets. No doubt, the
engineered models share a systems level approach for biological systems and elucidate an advantageous picture of cell
cycle framework and predictions, but a comprehensive illustrative picture is still missing.
We also applied machine learning on computational modelling performed in last 2 decades. The data revealed that
Li et al. (2004) were pioneers among others for modelling the complete yeast cell cycle from regulatory proteins [33].
They investigated the global dynamicity and cell cycle proliferation drivers’ trajectory and concluded that the network
is robustly stable for conducting its functions and declared G1 as a global attractor from simulation dynamics, and that
it is stable against all perturbations. Later, Davidich et al. (2008) presented a Boolean network model of the cell cycle
sequences which was solely based on the biochemical topology of the interactions reproducing the biological cell cycle
time sequence of protein activations. This minimalistic approach boosted the idea of predicting dynamical features of
proteins along with protein interaction networks of cell cycle. Furthermore, data suggested that in subsequent years,
Mangla et al. (2010) worked on synchronous models of the Budding and Fission yeast cell cycle and concluded that the
timing and robustness can be used as the basis for a testable hypothesis that could account for several needs-based
refinements in the model [33].
In addition to that Fauréet al. [34] were among the pioneers to model the mammalian cell cycle. They used the
most prominent drivers of the cell cycle machinery, i.e., Cyc D, E, A and B along with other regulators and effectors,
e.g., P27, Rb and E2F. Their synchronous and asynchronous Boolean modelling demonstrated the asymptotic behavior
of the regulatory proteins based on the experimental data and provided the biological justification for using multilevel
variables in future research.
Lately, Abroudi et al. (2017) published a paper on ‘A comprehensive complex systems approach to the study and
analysis of mammalian cell cycle control system in the presence of DNA damage stress’. They even considered G1-S
and G2-M checkpoints unlike Fauréet al. [34]. They first refined the published research on ODE mathematical models
and then included sub-systems, i.e., growth factors (GF), DNA damage, and G1/S and G2/M checkpoints. As advised
by Magla et al. [35], they applied a multi-level systems approach. The model was also used to assess the efficacy of

Copyright © 2021 MECS I.J. Education and Management Engineering, 2021, 2, 13-24
 The Cell Cycle Model: A Comprehensive Review and Extension Based on Machine Learning 21

DNA damage checkpoints in correctly arresting damaged cells and avoiding incorrect arrest of healthy cells and results
revealed 98.6% accuracy in correctly releasing healthy cells through checkpoints. Using ANN all of these models can
be computed completely and precisely in different layers, as shown in Fig. 10.

Fig. 10 Cell cycle modelling through ANN

Recently, Castro et al. (2019) developed an agent-based model of cell cycle adhering to the view that kinetic
parameters are a crucial aspect when studying reactions involving proteins in real cell cycle [36]. They compared the
results to a Boolean network model and found similar results in terms of following the correct sequence of phases. This
model could be a starting point for being used in the development of cancer drugs by adding cell cycle mutations that
match a specific type of tumor cell cycle and an agent representing the medication or treatment. Recently, Laomettachit
et al. [37] applied different mathematical modelling approaches including Boolean, discrete (stochastic), ODEs and
hybrids, as shown in Fig. 11, and concluded the same as previous in terms of cell robustness and stability, although
their models lacked the technical sophistications for mutated cell cycles.

Fig. 11 A cell cycle model by Laomettachit et al.

Copyright © 2021 MECS I.J. Education and Management Engineering, 2021, 2, 13-24
22 The Cell Cycle Model: A Comprehensive Review and Extension Based on Machine Learning

Scientists are working hard in order to model a cell cycle in a best possible way. As for now the models are quite
complicated. Microbiologist, botanists, and zoologists are working to model the cell cycle. Similarly, scientists in the
field of artificial intelligence are also trying to compare all models of cell cycles through machine learning [38-40]. The
use of innovative technology such as machine learning and Artificial Intelligence will help efficiently to model the cell
cycle using data of recent research [41,42].

5. Conclusion

The proliferative process of cell is important, and it has aided in research studies related to different areas of
biological sciences. Data derived through machine learning indicates that our current comprehensive understanding of
cell cycle is lagging and therefore, it requires the utilization of Artificial Intelligence in this domain so that
mathematical algorithms and machine learning techniques could be applied to sort the huge volumes of data. This study
was conducted with the same aim to demonstrate the role of different unconventional players that are actively
controlling cell cycle. It has been observed in our study that though a significant attention has been given to
conventional genomic regulators of cell cycle and their participation in different disease conditions have also been
evaluated, but a clear picture of mechanism of action of these unconventional players is still missing. Thus, we have
identified some key volume regulatory control elements in our study and presented the work that has been done on them
to better understand the process of cell cycle.
In order to efficiently evaluate our current progress in the field of cell biology, computational biologists are
working hard to identify different regulatory components of cell, its processes, machineries and cell cycle and
combining them to develop a computational model so it could aid in better understanding of the process along with
predicting the effects of perturbations that may be associated with any pathological condition. Different computational
tools were being used in these studies and computational models were built for different cellular processes of both
prokaryotic and eukaryotic cells. But our study through machine learning by using neural networks has found that
though, a lot of work has been done on cell and cell cycle, but we still lack an all-inclusive picture of cell cycle model.
Therefore, the unconventional players identified in our study would enable researchers to improve the current models of
cell cycle by incorporating more regulatory elements in them.
As previous computational models of cell cycle have not considered the importance of inclusion of volume
regulatory control elements in cell cycle, therefore, our work will provide a basis to improve the current knowledge
available on cell cycle through artificial intelligence. It would help in developing a comprehensive cell cycle model by
incorporating as many details as possible in the simplest possible way so that it could assist biologists in identifying and
evaluating the effects of perturbations in pathways. It will help in evaluating the role of various drugs on regulating
membrane potential, ionic homeostasis in the microenvironment and volume regulation in wet lab so that new targets
could be identified and their efficacy could be determined. Our study demonstrates that researchers need to emphasize
on volume regulatory machinery as well by identifying their mechanism of action and their utilization in therapeutics. It
would ultimately shift a focus from genome regulatory components to volume regulatory elements and if it will work
well, it would enable us to overcome the effect of Multi Drug Resistance (MDR). The jump from wet lab
experimentation to computational modelling has proven vital. There is now a need to take a leap towards machine
learning and artificial intelligence methodologies to better understand the working of the cell. Together, these
phenomena may pave the path for bringing innovation in the field of regenerative medicine and to develop new
therapeutic solutions to prevent or restore disease states such as uncontrolled cell proliferation in cancer.

References

[1] Alberts, B., Johnson, A., Lewis, J., Raff, M., Roberts, K., & Walter, P. (2002). Molecular Biology of the Cell 4th edn (New
York: Garland Science). Ann Bot, 91, 401.
[2] Gali-Muhtasib, H., & Bakkar, N. (2002). Modulating cell cycle: current applications and prospects for future drug
development. Current cancer drug targets, 2(4), 309-336.
[3] Norbury, C., & Nurse, P. (1992). Animal cell cycles and their control. Annual review of biochemistry, 61(1), 441-468.
[4] Lodish, H., Darnell, J. E., Berk, A., Kaiser, C. A., Krieger, M., Scott, M. P. & Matsudaira, P. (2008). Molecular cell biology.
Macmillan.
[5] Patt, H. M., & Quastler, H. (1963). Radiation effects on cell renewal and related systems. Physiological reviews, 43(3), 357-
396.
[6] Howard, A., c. Heredity suppl, 1953. 6: p. 261-273.
[7] Blagosklonny, M. V., & Pardee, A. B. (2002). The restriction point of the cell cycle. Cell cycle, 1(2), 102-109.
[8] Munawar HS, Awan AA, Khalid U, Munawar S, Maqsood A. Revolutionizing Telemedicine by Instilling H. 265. International
Journal of Image, Graphics & Signal Processing. 2017 May 1;9(5).
[9] Munawar, H. S., Qayyum, S., Ullah, F., & Sepasgozar, SBig Data and Its Applications in Smart Real Estate and the Disaster
Management Life Cycle: A Systematic Analysis. Big Data and Cognitive Computing; 2020, 4(2).
[10] Munawar, H. S., Zhang, J., Li, H., Mo, D., & Chang, L. Mining multispectral aerial images for automatic detection of strategic
bridge locations for disaster relief missions. In Pacific-Asia Conference on Knowledge Discovery and Data Mining (pp. 189-
200). Springer, Cham. . (2019, April).

Copyright © 2021 MECS I.J. Education and Management Engineering, 2021, 2, 13-24
 The Cell Cycle Model: A Comprehensive Review and Extension Based on Machine Learning 23

[11] Munawar HS, Hammad A, Ullah F, Ali TH. After the flood: A novel application of image processing and machine learning for
post-flood disaster management. InProceedings of the 2nd International Conference on Sustainable Development in Civil
Engineering (ICSDC 2019), Jamshoro Pakistan 2019 Dec (pp. 5-7).
[12] Munawar, H. S. Applications of Leaky-wave Antennas: A Review. International Journal of Wireless and Microwave
Technologies (IJWMT); 2020, 10(4), 56-62.
[13] Munawar, H. S. An Overview of Reconfigurable Antennas for Wireless Body Area Networks and Possible Future Prospects.
International Journal of Wireless and Microwave Technologies (IJWMT);2020, 4, 1-8.
[14] Munawar, H. S., Khalid, U., Jilani, R., & Maqsood, A. (2017). Version Management by Time Based Approach in Modern
Era. International Journal of Education and Management Engineering, 4, 13-20.
[15] Munawar, H. S. Reconfigurable Origami Antennas: A Review of the Existing Technology and its Future Prospects.
International Journal of Wireless and Microwave Technologies (IJWMT); 2020, 4, 34-38.
[16] Munawar, H. S., & Maqsood, A. Isotropic Surround Suppression based Linear Target Detection using Hough Transform.
[17] Pledger, W. J., Stiles, C. D., Antoniades, H. N., & Scher, C. D. (1977). Induction of DNA synthesis in BALB/c 3T3 cells by
serum components: reevaluation of the commitment process. Proceedings of the National Academy of Sciences, 74(10), 4481-
4485.
[18] Campisi, J. U. D. I. T. H., & Pardee, A. B. (1984). Post-transcriptional control of the onset of DNA synthesis by an insulin-like
growth factor. Molecular and cellular biology, 4(9), 1807-1814.
[19] Rossow, P. W., Riddle, V. G., & Pardee, A. B. (1979). Synthesis of labile, serum-dependent protein in early G1 controls animal
cell growth. Proceedings of the National Academy of Sciences, 76(9), 4446-4450.
[20] Bartek, J., & Lukas, J. (2001). Pathways governing G1/S transition and their response to DNA damage. FEBS letters, 490(3),
117-122.
[21] Massagué, J. (2004). G1 cell-cycle control and cancer. Nature, 432(7015), 298.
[22] Grissmer, S., Nguyen, A. N., & Cahalan, M. D. (1993). Calcium-activated potassium channels in resting and activated human T
lymphocytes. Expression levels, calcium dependence, ion selectivity, and pharmacology. The Journal of general physiology,
102(4), 601-630.
[23] Lang, F., Busch, G. L., Ritter, M., Volkl, H., Waldegger, S., Gulbins, E., & Haussinger, D. (1998). Functional significance of
cell volume regulatory mechanisms. Physiological reviews, 78(1), 247-306.
[24] Lin, C. S., Boltz, R. C., Blake, J. T., Nguyen, M., Talento, A., Fischer, P. A., ... & Garcia, M. L. (1993). Voltage-gated
potassium channels regulate calcium-dependent pathways involved in human T lymphocyte activation. Journal of Experimental
Medicine, 177(3), 637-645.
[25] Lang, F. (2007). Mechanisms and significance of cell volume regulation. Journal of the American college of nutrition, 26(sup5),
613S-623S.
[26] Sanchez, Y., Wong, C., Thoma, R. S., Richman, R., Wu, Z., Piwnica-Worms, H., & Elledge, S. J. (1997). Conservation of the
Chk1 checkpoint pathway in mammals: linkage of DNA damage to Cdk regulation through Cdc25. Science, 277(5331), 1497-
1501.
[27] Kahl, C. R., & Means, A. R. (2003). Regulation of cell cycle progression by calcium/calmodulin-dependent
pathways. Endocrine reviews, 24(6), 719-736.
[28] Devault, A., Cavadore, J. C., Fesquet, D., Labbé, J. C., Lorca, T., Picard, A., ... & Doree, M. (1991, January). Concerted roles
of cyclin A, cdc25+ mitotic inducer, and type 2A phosphatase in activating the cyclin B/cdc2 protein kinase at the G2/M phase
transition. In Cold Spring Harbor symposia on quantitative biology (Vol. 56, pp. 503-513). Cold Spring Harbor Laboratory
Press.
[29] Takuwa, N., Zhou, W., Kumada, M., & Takuwa, Y. (1993). Ca (2+)-dependent stimulation of retinoblastoma gene product
phosphorylation and p34cdc2 kinase activation in serum-stimulated human fibroblasts. Journal of Biological Chemistry, 268(1),
138-145.
[30] Diehl, J. A. (2002). Cycling to cancer with cyclin D1. Cancer biology & therapy, 1(3), 226-231.
[31] Se, V., Rajala, N., Spiller, D., & White, M. (2004). Calcium-dependent regulation of the cell cycle via a novel MAPK–NF-κB
pathway in Swiss 3T3 cellsAbbreviations used in this paper. The Journal of Cell Biology, 166(5), 661-672.
[32] Ohta, Y., Ohba, T., & Miyamoto, E. (1990). Ca2+/calmodulin-dependent protein kinase II: localization in the interphase
nucleus and the mitotic apparatus of mammalian cells. Proceedings of the National Academy of Sciences, 87(14), 5341-5345.
[33] Davidich, M. I., & Bornholdt, S. (2008). Boolean network model predicts cell cycle sequence of fission yeast. PloS one, 3(2),
e1672.
[34] Fauré, A., Naldi, A., Chaouiya, C., & Thieffry, D. (2006). Dynamical analysis of a generic Boolean model for the control of the
mammalian cell cycle. Bioinformatics, 22(14), e124-e131.
[35] Mangla, K., Dill, D. L., & Horowitz, M. A. (2010). Timing robustness in the budding and fission yeast cell cycles. PLoS
One, 5(2), e8906.
[36] Castro, C., Flores, D. L., Vargas, E., Cervantes, D., & Delgado, E. (2019). Agent-Based Model of the Budding Yeast Cell
Cycle Regulatory Network. In World Congress on Medical Physics and Biomedical Engineering 2018 (pp. 531-534). Springer,
Singapore.
[37] Laomettachit, T., Chen, K. C., Baumann, W. T., & Tyson, J. J. (2016). A model of yeast cell-cycle regulation based on a
standard component modeling strategy for protein regulatory networks. PloS one, 11(5), e0153738.
[38] Munawar, H. S., Khalid, U., & Maqsood, A. Fire detection through Image Processing; A brief overview.
[39] Munawar, H. S., Khalid, U., & Maqsood, A. Modern day detection of Mines; Using the Vehicle Based Detection Robot.
[40] Munawar HS, Awan AA, Maqsood A, Khalid U. REINVENTING RADIOLOGY IN MODERN ERA.
[41] Munawar, H. S. (2020). Flood Disaster Management: Risks, Technologies, and Future Directions. Machine Vision Inspection
Systems: Image Processing, Concepts, Methodologies and Applications, 1, 115-146.
[42] Munawar, H. S. (2020). Image and Video Processing for Defect Detection in Key Infrastructure. Machine Vision Inspection
Systems: Image Processing, Concepts, Methodologies and Applications, 1, 159-177.

Copyright © 2021 MECS I.J. Education and Management Engineering, 2021, 2, 13-24
 24 The Cell Cycle Model: A Comprehensive Review and Extension Based on Machine Learning

Authors’ Profiles

Mustafa Pasha is an independent consultant on computational modelling and simulation on medical and health
related topics. He has a master’s in computational sciences and engineering from NUST, Pakistan and a PhD in
Applied computing from Lincoln, New Zealand. He has a dedicated set of expertise in drug design and discovery,
his past work includes work on cancer cell proliferation and human cell cycle modelling and simulation. Besides
his research profile, he has number of achievements in health business, procurement, and novel solutions
consultations. He holds the privilege to be nominated in Canterbury Business Champion, New Zealand. His
research interests include, Pharmaceutical Formulations, Intelligent Modelling and Simulation, Artificial
intelligence, Machine learning, Data Analysis, Industrial Business Consultancy and regulatory affairs.

Khurram Munawar is a PhD student at the Lincoln University (New Zealand) He is a multi-disciplinary
researcher with experience in Computer Visualization, machine learning, computational sciences, computational
modeling and artificial intelligence, he has several international publications in various journals and conferences
and has actively been working in Visualization and Artificial Intelligence domain.

Asma Talib Qureshi is a MS student in the Healthcare Biotechnology discipline at NUST, Pakistan. She has
also been a member a Pakistan Society of Basic and Applied Neurosciences. She has expertise in biotechnology,
cancer research and neurosciences. She has worked on autoimmune and viral diseases and currently her main
area of research interest is cancer biology and nervous system disorders.

How to cite this paper: Mustafa Kamal Pasha, Khurram Munawar, Asma Talib Qureshi, " The Cell Cycle Model: A Comprehensive
Review and Extension Based on Machine Learning", International Journal of Education and Management Engineering (IJEME),
Vol.11, No.2, pp. 13-24, 2021. DOI: 10.5815/ijeme.2021.02.02

Copyright © 2021 MECS I.J. Education and Management Engineering, 2021, 2, 13-24

View publication stats