partitioning, and identification of abnormalities (Bir & Balas, 2020).

CNNs
have proven effective in accurately categorizing lung nodules in CT scans,
hence improving the ability to detect lung cancer at an early stage (Xu et
al., 2021). Utilizing CNNs in this field has led to increased precision rates
in comparison to conventional techniques, thereby minimizing incorrect
identifications and enhancing diagnostic certainty (Bir & Balas, 2020;
Sarvamangala & Kulkarni, 2022). In addition, CNNs have been utilized to
segment brain tumors in MRI scans, enabling accurate identification and
classification of tumors. This capability is essential for the development of
treatment strategies. In addition, CNNs have been utilized to detect diabetic
retinopathy from retinal pictures. They are capable of accurately classifying
the severity of retinopathy, which enables early intervention and effective
disease management(Ghosal et al., 2019).

9.4.3 Recurrent Neural Networks in Healthcare Data

RNNs are specifically intended to process sequential data, which makes
them well-suited for evaluating time-series data in the healthcare field
(Lalapura et al., 2021). RNNs have different versions, one of which is long
short-term memory (LSTM) networks. These networks are very good at
collecting long-term dependencies, which are crucial for various healthcare
applications (Y. Chen & Li, 2021; Neves et al., 2021). An important use of
RNNs is to forecast patient outcomes using EHRs. RNNs can forecast
forthcoming health occurrences, such as hospital readmissions or the
initiation of chronic illnesses, by examining sequences of medical events
and patient history (Ho et al., 2021). This assists in proactive healthcare
management. In addition, RNNs have been employed to diagnose diseases,
like as heart failure, at an early stage (Ledbetter et al., 2021). This is
achieved by continuously monitoring changes in patient data over time,
which allows for prompt interventions and has the potential to enhance
patient outcomes. RNNs are utilized in healthcare for various NLP
activities, including the anonymization of patient notes and the extraction of
significant information from clinical narratives (Ackerson et al., 2021). This
program facilitates enhanced data management and enables more informed
clinical decision-making (Ackerson et al., 2021; A. Rahman et al., 2021).

9.5 ADVANCED AI APPLICATIONS IN

BIOINFORMATICS
9.5.1 Natural Language Processing and Genomics

9.5.1.1 NLP for Biological Text Mining

NLP is becoming an essential technology in bioinformatics for extracting
relevant information from the extensive collection of biological literature.
Due to the rapid increase in scientific publications, researchers are able to
automate the extraction and analysis of biological data using NLP
approaches. For instance, NLP models like PubTator Central and BioBERT
have been created to annotate biomedical texts and extract significant
biological elements and relationships. NLP approaches are widely
employed for extracting associations between genes and diseases. NLP
systems can analyze scientific literature to detect and emphasize references
to genes and diseases, as well as their connections. The implementation of
this automated procedure has greatly expedited the detection of genetic
indicators for different illnesses, enhancing the productivity and efficacy of
biomedical investigation. The integration of NLP with genomic data
improves the comprehension of genetic functions and disease mechanisms
by establishing connections between textual information extracted from
scientific literature and genomic datasets. The integration of clinical notes
with genomic data in EHRs is particularly beneficial for identifying genetic
variants that are related to specific clinical characteristics. For example,
NLP approaches have been employed to remove identifying information
from patient notes and extract clinical data, which is subsequently
connected to genomic data for comprehensive study.
This systematic approach is also advantageous for annotating genes
using information derived from books. Researchers can enhance the
functional annotation of genes by merging data from sources such as
Ensembl and NCBI. This can assist in identifying gene regulation
mechanisms and possible targets for therapeutic interventions.

9.5.2 AI in Genomic Research

9.5.2.1 AI in DNA Sequencing

AI has greatly improved the process of DNA sequencing by enhancing the
precision and efficiency of sequence analysis (Mumtaz et al., 2023).
DeepVariant, a DL model created by Google, demonstrates this by utilizing
CNNs to enhance the precision of variant calling from sequencing data
(Afanasiev et al., 2021; Giudice et al., 2021). This model can differentiate
between authentic genetic variations and errors that occur during the
sequencing process, resulting in more dependable sequencing outcomes.
In addition, AI has been utilized in nanopore sequencing, a technology
renowned for its ability to generate data in real time but with significant
mistake rates (Jena & Pathak, 2023). AI models, which have been taught
using large sequencing datasets, can promptly rectify mistakes, hence
enhancing the suitability of nanopore sequencing for clinical purposes.
These findings have substantial ramifications for areas such as customized
treatment and genetic research (Kacew et al., 2021).

9.5.2.2 Predictive Modeling and Success Stories

AI-driven predictive modeling has emerged as a fundamental aspect of
genomic research, facilitating the anticipation of disease vulnerability and
treatment outcomes through the analysis of genetic information (Kosvyra et
al., 2020). ML methods, such as Random Forests and Gradient Boosting
Machines, combine diverse biological data, including genomic, epigenomic,
and transcriptome information, to generate precise predictions on illness
outcomes (Srinivasu et al., 2022). An exemplary achievement in AI-
powered genomic research is the application of predictive modeling in the
field of cancer genomics. AI models have been employed to detect
mutations in the BRCA1 and BRCA2 genes, which are linked to a
heightened susceptibility to breast and ovarian malignancies (Wang et al.,
2020). These models enable prompt identification and individualized
treatment strategies, leading to substantial enhancements in patient results
(Du et al., 2023).
Furthermore, AI has been influential in finding genetic variations linked
to uncommon medical conditions. Through the analysis of extensive
genomic datasets, AI algorithms can identify uncommon genetic variations
that conventional methods may overlook (Robson & Ioannidis, 2023). This
ability assists in the identification and management of rare genetic illnesses.
DL methods have enhanced the identification of harmful variations in genes
linked to neurodevelopmental disorders, thereby offering a fresh
understanding of the genetic foundation of these problems (Haga et al.,
2020; Stein et al., 2022).

9.5.2.3 Hereditary Fructose Intolerance

Hereditary Fructose Intolerance (HFI) is an uncommon genetic condition
resulting from mutations in the ALDOB gene, responsible for producing the
enzyme aldolase B (Coffee et al., 2010). This enzyme is essential for the
process of fructose metabolism, and if there is a lack of it, harmful
metabolites will build up in the body when fructose is consumed. The
utilization of AI and ML has been employed to augment the diagnosis and
comprehension of HFI through the analysis of genetic data and the
prediction of the consequences of specific mutations (Ferri et al., 2012).
The latest progress in AI has made it possible to identify harmful variations
in the ALDOB gene by examining sequencing data (Esposito et al., 2004).
The utilization of AI-driven methods has significantly enhanced the
precision of genetic testing for HFI, facilitating prompt detection and more
effective treatment of the illness. For example, predictive models can
anticipate the probable consequences of particular ALDOB mutations,
aiding clinicians in comprehending the extent of the condition in individual
patients (Adamowicz et al., 2007).
Moreover, the integration of AI with EHRs enables the detection of
patients displaying symptoms that suggest HFI, even in cases when genetic
testing is not readily accessible (Beyzaei et al., 2023). Natural language
processing (NLP) methods can retrieve pertinent clinical data from patient
records, thereby aiding in the prompt identification and treatment of HFI
(Ferri et al., 2012; Gunduz et al., 2021).

9.6 AI IN HEALTHCARE AND PUBLIC HEALTH

AI is transforming healthcare and public health through its sophisticated
powers in data processing, pattern recognition, and predictive analysis. This
transformation has a significant impact on precision medicine and
personalized healthcare. AI plays a crucial role in tailoring medicines
according to individual genetic profiles and improving public health
management through predictive analytics.

9.6.1 Precision Medicine and Personalized Healthcare

Precision medicine, along with personalized healthcare, aims to customize
medical treatments based on individual patient characteristics, in contrast to
conventional techniques that are designed to fit all patients uniformly (Sisk
et al., 2020). AI is essential in facilitating this transition by allowing for the
incorporation and examination of intricate data to enhance patient results
and treatment approaches.

9.6.1.1 AI’s Role in Precision Medicine

AI is becoming essential in precision medicine, specifically in genomics
and the management of genetic illnesses. ML algorithms have the capability
to analyze large quantities of genomic data in order to find genetic changes
that are associated with diseases (Takei et al., 2018). This ability helps in
the advancement of targeted medicines. AI systems excel in identifying
mutations linked to particular types of cancer, facilitating timely detection
and tailored treatment strategies. This is particularly apparent in the field of
oncology, where AI assists in customizing treatments according to the
genetic characteristics of tumors, hence improving effectiveness and
reducing adverse effects (Girotti et al., 2016; Kurnaz & Loaiza-Bonilla,
2019). AI plays a crucial role in discovering biomarkers for different
genetic illnesses. Biomarkers play a vital role in the diagnosis and
monitoring of diseases such as Alzheimer’s, Parkinson’s, and metabolic
disorders (Bahado-Singh et al., 2022).
AI has been employed to detect genetic changes in individuals diagnosed
with thalassemia, a blood disorder resulting from abnormalities in the HBB
gene for beta-thalassemia and the HBA1/HBA2 genes for alpha-thalassemia
(Takei et al., 2018). This level of accuracy enables more efficient
administration and therapy of the illness. AI can utilize genetic data to
forecast patient reactions to dietary fructose in cases of HFI. HFI is a result
of genetic abnormalities in the ALDOB gene, which causes a shortage of
aldolase B (Petersson et al., 2023). AI can assist in the early detection of
people with this ailment, enabling dietary adjustments that can prevent the
occurrence of severe symptoms such as liver and kidney damage (Sisk et
al., 2020).

9.6.1.2 Developing Personalized Treatment Plans

Personalized treatment plans require the integration of data from multiple
sources, such as lifestyle data and genetic test results (Güvenç Paltun et al.,
2021). AI systems evaluate this data to uncover patterns and connections
that provide insights for personalized treatment plans. Pharmacogenomics is
a notable field where AI is used to forecast the effectiveness of drugs and
possible adverse effects by analyzing a patient’s genetic makeup (Gkouvas,
2022). This approach minimizes the need for trial and error when
administering pharmaceuticals. AI can assist in customizing blood
transfusion schedules and chelation therapy for individual patients with
hereditary illnesses such as thalassemia (Spaulding & Deogun, 2011). This
individualized strategy enhances patient outcomes and enhances quality of
life (Gifari et al., 2021). AI can provide dietary recommendations for
genetic fructose intolerance that eliminates fructose, sucrose, and sorbitol.
This helps prevent severe metabolic crises and improves the management of
the condition (Rezayi et al., 2022).