Page 50 - AN-2-4

Advanced Neurology Genomic insights into Alzheimer

the initial or refined model exhibited superior quality. matrix (QM) methods were used for predicting the toxicity
Conversely, the overall model quality, represented by a of each B- and T-cell epitope.
z-score, was calculated based on the model’s similarity to
other experimentally valid structures. Visual examination 2.8. Retrieval of genetic properties for each APP
and consolidation were performed based on the negative mutation
state of each Z-score . Notably, all models demonstrated Data concerning the position and codon changes for each
[33]
higher average quality scores in the refined structures APP mutation were collated from the AlzForum database.
compared to the initial ones. Furthermore, all mutations This database provided comprehensive information on
demonstrated comparable results in the refined structures chromosomal position, coding or non-coding properties,
in terms of quality. mutation type, codon and amino acid alterations, reference
[21]
The protein model visualization software PyMOL isoform, and genomic region for each mutation . In
(software version 2.5.4) was employed to visualize and addition, more detailed information about each amino
superimpose the 3D structures of both the wild-type acid related to APP mutations was gathered, encompassing
and mutated human APP protein . The degree of the codon sequence, R group charge, and molecular mass.
[34]
differentiation between superimposed protein models was
quantified using the root mean square deviation (RMSD) 2.9. Literature review for the identification of APP
score . mutation-related clinical phenotype, pathogenicity,
[34]
and neuropathology
2.5. Identification of T-cell epitopes in the wild-type An extensive literature review was conducted to elucidate
and mutated human APP sequences the clinical and neuropathological phenotypes associated
The NetCTL v1.2 server (https://services.healthtech.dtu. with each genetic subtype of fAD. The initial phase of
dk/service.php?NetCTL-1.2) served as the tool for the literature review involved the utilization of diverse
[35]
predicting cytotoxic T lymphocyte (CTL) epitopes within conjunctions of keywords and Boolean operators such
protein sequences, integrating peptide prediction through as “AND,” “OR,” and “NOT” on the PubMed (pubmed.
MHC Class 1 binding, proteasomal C terminal cleavage, ncbi.nlm.nih.gov) and Embase (embase.com) databases.
and transporter associated with antigen processing (TAP) The comprehensive list of keywords and terms employed
transport efficiency. In this study, parameters for identifying in the search included: “Alzheimer’s Disease,” “familial
T-cell epitopes from both wild-type and mutated human AD,” “amyloid precursor protein,” “clinical phenotype,”
APP sequences included a weight of 0.15 on C terminal “clinical symptoms,” “neuropathology,” and “ethno-genetic
cleavage, a weight of 0.05 on TAP transport efficiency, subtype.” This literature search was conducted using various
and a threshold for epitope identification set at 0.75. combinations of keywords and Boolean operators across
Finally, epitopes were predicted and collected based on the both databases, encompassing relevant abbreviations and
combined score generated by the NetCTL v1.2 server. truncations (*).

2.6. Identification of the linear and conformational After the initial review, relevant data were selectively
B-cell epitopes in the wild-type and mutated human chosen based on research quality and validity. Subsequently,
APP sequences a list of mutations on the APP gene associated with fAD
was created, drawing information from the ALZForum
Both linear and conformational B-cell epitopes were database. The previously mentioned keywords and terms
predicted from the structures of both the wild-type were then combined with mutation subtype descriptions,
and mutated human APP sequences using the Ellipro
server (http://tools.iedb.org/ellipro/) under the IEDB which included ethno-genetic names (e.g., “Flemish” and
[36]
database (https://www.iedb.org/) . Finally, the BepiPred “E693G”). This combination aimed to identify all relevant
[37]
v2.0 server (https://services.healthtech.dtu.dk/service. mutation-specific research papers. The searches were
[38]
php?BepiPred-2.0) was used to predict linear B-cell conducted through Google Scholar, as well as Medline
and Cochrane Library, to augment the volume of results,
epitopes from the linear sequences, providing validation
for the linear B-cell epitopes predicted by the Ellipro server. particularly given the limited caseload available within the
literature for rarer subtypes. In addition, further relevant
2.7. Toxicity prediction of B-cell and T-cell epitopes research was identified by exploring the references of
relevant systematic reviews and examining articles listed
The toxicity of the predicted T-cell epitopes and the linear in the ALZForum database.
B-cell epitopes was assessed using the ToxinPred server
(http://crdd.osdd.net/raghava/toxinpred/) [39,40] . In this The data collated from the extensive literature search
study, both support vector machines (SVM) and quantitative underwent a quality review based on the representativeness

Volume 2 Issue 4 (2023) 5 https://doi.org/10.36922/an.1734