Page 30 - GPD-2-3

Gene & Protein in Disease Signatures construction strategies for TC

respectively estimated to define each signature, taking
BRAFV600E mutation into account. Thus, K-M survival
analysis showed that patients with high glycolysis or GLUT
scores possessed a worse prognosis in PDTC/ATC .
[50]
3.3.4.2. Signatures related to alternative splicing events
As reported, splicing events significantly influence
the occurrence of cancer [71-74] . Lin et al. obtained data
from 496 patients with AS profiles derived from TCGA
SpliceSeq and clinical data. Lasso regression analysis was
performed to develop seven types (alternate acceptor site
[AA], alternate promoter [AP], alternate terminator [AT],
alternate donor site [AD], exon skip [ES], retained intron
[RI], and mutually exclusive exons [ME]) of splicing events.
Moreover, 20 alternative splicing (AS) events were screened
to recognize the most significant prognostic signature. This
signature was then verified as an independent predictor
by the ROC curve (AUC value is 0.843) and multivariate
analysis . Furthermore, Han et al. have also concerned
[52]
with the pivotal role of the alternative splicing events in
TC prognosis and built a prediction model including
AA, ES, AD, ME, and AT events with accurate prognostic
efficacy , indicating that the prognosis effect of alternative
[51]
splicing events on TC was a promising research subject.
Figure 3. The flow chart of strategy 1: Signatures based on differentially
expressed mRNAs/lncRNAs/miRNAs 3.3.4.3. Signatures correlated with genetic alterations
Recently, prognostic biomarkers with genetic alterations in
various tumors have become a research focus . Gandolfi
[75]
et al. constructed a genetic alteration-related signature that
[54]
consisted of three genetic variables . The first step of their
study was the acquisition of clinical data on PTC patients.
A series of analyses of the genetic profiles were then
conducted in distant metastasis (DM) PTCs and control
samples to obtain the differential alterations between
DMs and controls. As a result, three genetic variables
including duplication of Chr1q, duplication of Chr5p
harboring TERT locus, and mutations in TERT promoter
displayed strong relevance with distant metastasis. Hence,
a distinctive signature integrated with three genetic
features was identified as Thyroid TERT Chr1q (THYT1).
K-M survival analysis was then performed to assess the
association of the THYT1 signature with the progression
of distant metastasis. Furthermore, through the univariate
and multivariate Cox models, the THYT1 signature was
demonstrated to be an independent risk factor that was
capable of predicting the aggressiveness in PTCs.
Figure 4. The flow chart of strategy 2: Signatures based on DEGs
with specific biological functions. Verification of biomarker* means 3.3.4.4. Signatures constructed with genetic mutation
the validation methods include survival curve, ROC curve, and Cox
regression analysis. Genetic mutation is a quite common phenomenon in
TC [76,77] . Han et al. downloaded mutation data from
glycolysis signature using microarray data of PDTC and 487 samples, and expression profiling data from 502
[50]
ATC patients. The glycolysis and GLUT scores were then PTC and 58 normal samples. These samples were then

Volume 2 Issue 3 (2023) 14 https://doi.org/10.36922/gpd.1138