Page 116 - Read Online

Ratnapriya. J Transl Genet Genom 2022;6:240-256 https://dx.doi.org/10.20517/jtgg.2021.54 Page 250

alterations (single nucleotide polymorphisms, deletions or insertions, haplotypes, copy number variations)
or RNA modification, whereas RNA characteristics can include RNA sequences, expression levels, RNA
processing, and circulating RNA such as micro RNAs (miRNAs) and long-noncoding RNAs (lncRNAs).

The main aim of GWAS is biological discovery and gaining insights into the genes and pathways involved
in AMD. However, in the light of immense success on AMD-GWAS, the development of DNA biomarkers
holds great promises for personalized risk prediction and early detection. Several studies have looked into
the utility of risk variants in AMD risk prediction and disease progression with varying success (reviewed in
Refs. [111,112] ). These could be further improved by integrating information from all risk variants, rare variants
as well as molecular risk factors. Additionally, as AMD risk is conferred by multiple genetic variants,
accessing the cumulative impact of risk allelesthrough polygenic risk scores (PRS) is likely to have better
[113]
success for risk prediction . PRS are the weighted sum of the individual effects of many risk alleles, with
the weight based on the effect size estimated from GWAS and optionally re-estimated to account for other
properties such as LD. Using data from two large clinical trials, Age-Related Eye Disease Study (AREDS)
and AREDS2, it was shown that the addition of PRS to the demographic/environmental risk factors
[114]
considerably improved the prediction performance . A more recent study combining deep learning with
survival analysis achieved high prognostic accuracy in predictions of progression to late AMD [115,116] .

Despite analytical and clinical validation, a biomarker should demonstrate a clear effect in improving
patient management, and an added value to the available treatment and management options for decision-
making for AMD patients. However, at this stage gene therapy is not available for AMD and no direct
benefits for patients have been demonstrated by identifying genetic risk factors. Thus, the American
Academy of Ophthalmology currently does not recommend genetic testing for AMD patients. As future
studies will shed more light on our understanding of disease etiology associated with AMD risk variants that
show the value of treatment tailored to individuals’ genetic risk, genetic testing for AMD may become a
routine in clinical practices.

There are even fewer examples of developing prognostic and predictive biomarkers based on gene
expression signature. Several issues continue to impede progress in the clinic. One key limitation is the lack
of transcriptome profiles available for AMD samples. Secondly, even when such studies are done, the
sample sizes are small and a little overlap is reported when comparing gene signatures from different
studies. Most studies also use convenience samples of heterogeneous patients for obtaining gene expression
signatures, which makes it harder to replicate. A special class of RNA, called microRNA, is a small non-
coding RNA molecule, thatis secreted in the circulation and exists in a stable form. As they are involved in
post-transcriptional gene regulation, they have been good candidates for AMD biomarker discovery in
systemic and ocular fluids. Multiple studies have performed comparative analyses of controls and AMD
biofluids in search of biomarkers and have shown exciting possibilities, but the considerable inconsistency
[110]
between different studies remains a major challenge for their clinical usefulness (reviewed Ref. ).
Biomarker identification requires a greater understanding of disease pathobiology. Thus, performing the
transcriptome studies in the AMD-relevant tissues (RPE, retina and choroid) are likely to aid in biomarker
discovery. However, gene signatures developed on these tissues can have limited application because of their
inability to screen patients noninvasively. Thus, accessing the predictive capability of genes identified
through genomic and molecular basis in circulating systemic and ocular fluids can have greater clinical
potential. For example, VEGF and components of the complement system have been extensively
investigated in AMD patients using metabolomic and proteomic approaches. Analysis of VEGF and its
receptor in aqueous humor and plasma in various studies have failed to find a consensus change in VEGF