Page 115 - Read Online

Page 249 Ratnapriya. J Transl Genet Genom 2022;6:240-256 https://dx.doi.org/10.20517/jtgg.2021.54

findings. There are several methods for predicting the association between gene expression and the trait that
leverages on eQTL reference, individual or summary-level GWAS data, and LD information.
[100]
Transcriptome-Wide Association Studies (TWASs) and Summary-data based Mendelian Randomization
[101]
(SMR) are two of the most commonly used methods for achieving this goal.

TWAS is a test for significant association between the cis component of gene expression and the GWAS
[100]
trait that can help in identifying the target genes . To date, two studies have looked into the association
between the genetic component of gene expression and AMD. We showed that integrating GWAS
[19]
summary level data with eQTL generated for ~500 retinas led to the identification of 61 transcriptome-
wide significant gene-AMD associations. Of these, 38 genes were present within 1 Mb of 14 AMD-GWAS
loci and 23 genes outside the GWAS loci. This analysis helped in locating the target genes at the known loci
[80]
as well as identifying RLBP1, PARP12 and HIC1 as likely new AMD-associated genes . The second study
[19]
analyzed the same GWAS data with eQTL information from 27 different human tissues and reported 106
genes significantly associated with AMD variants in at least one tissue. Among these, 54 genes were
significantly AMD-associated in one or more tissues and 16 genes (ADAM19, ARMS2, BTBD16, CFH,
a
CFHR1, CFHR3, GPR108, PILRA, PILRB, PLA2G12A, PLEKHA1, PMS2P1, PPIL3, RDH5, STAG3L5P, n d
TNFRSF10A) were associated with AMD disease status in over 10 tissues . These studies highlight the
[85]
application of integrative analysis in biological insights.

MOVING FROM TISSUE TO CELL LEVEL RESOLUTION
The majority of approaches for identifying genes and pathways in complex diseases including AMD have
used transcriptome profiles from the bulk tissues. However, tissues are made from many different cell types
and bulk expression levels represent the average signals from multiple cell types present in that tissue. As
discussed above, scRNA-seq presents a good alternative moving forward. However, currently we have
limited single-cell data available for ocular tissues. Thus, the majority of studies have focused on identifying
the cell types that are relevant in AMD pathology by looking at individual gene expression [57,86,102] or by
performing cell-type-specific enrichment of AMD-associated genes . Going onward, single-cell eQTL
[103]
analysis will be critical for defining the precise cellular context for disease-associated risk variants and their
target genes . However, single-cell eQTL analysis has less discovery power, and profiling in a large cohort
[104]
needed for such studies is still cost-prohibitive. Thus, a combination of bulk and single-cell analyses
combined with deconvolution approaches can help in understanding the cellular context for disease-
associated variants and their target genes. Finally, the construction of gene-regulatory networks at the
single-cell level can help in unraveling the cell-type-specificity of gene interactions that leads to the disease.

BIOMARKER DISCOVERY IN AMD
A biomarker can be a substance or structure measured in body parts, fluids, or products that can affect or
predict disease incidence, enable early detection of disease, and improve diagnostic classification to better
inform individualized treatment. There have been several efforts to profile AMD-specific genomic,
proteomic, metabolomic, and imaging-based biomarkers. Among these, studies involving systemic and
ocular fluids (urine, tears, serum, and plasma) have achieved moderate success and implicated several
components of the immune and complement system as potential biomarkers for AMD. Additionally,
imaging-based structural and functional biomarkers have also shown great potential in predicting disease
progression [105,106] and treatment outcomes . They have the added advantage of being less invasive. Several
[107]
excellent and comprehensive overview of all potential biomarkers and their applicability in AMD has been
published [108-110] . Thus, here we focus our discussion on genomic biomarkers, which are measurable DNA
and/or RNA characteristics that can be used as an indicator of normal biologic processes, pathogenic
processes, and/or response to therapeutic or other interventions. These characteristics could be DNA