In Silico Analysis of Post-COVID-19 Condition (PCC) Associated SNP rs9367106 Predicts the Molecular Basis of Abnormalities in the Lungs and Brain Functions

TL;DR

This study uses computational methods to explore how a genetic variant linked to long-term effects of COVID-19 may affect lung and brain functions.

Contribution

The study provides a novel in silico analysis of the molecular mechanisms of a PCC-associated SNP and its impact on gene regulation.

Findings

The SNP rs9367106 is located near and interacts with the FOXP4 gene and its antisense RNA, suggesting regulatory roles.

The G>C mutation alters RNA structure and transcription factor binding, potentially affecting gene expression in lung tissue.

The SNP is linked to brain-related gene MED20, which may explain neurological symptoms in PCC.

Abstract

Long- or post-COVID-19 syndrome, which is also designated by WHO as Post COVID-19 Condition (PCC), is characterized by the persistent symptoms that remain after recovery from SARS-CoV-2 infection. A worldwide consortium of Long COVID-19 Host Genetics Initiative (Long COVID-19 HGI) identified an SNP rs9367106 (G>C; chr6:41,515,652, GRCh38, p = 1.76 × 10−10, OR = 1.63, 95% CI: 1.40–1.89) that is associated with PCC. Unraveling the functional significance of this SNP is of prime importance to understanding the development of the PCC phenotypes and their therapy. Here, in Silico, I explored how the risk allele of this SNP alters the functional mechanisms and molecular pathways leading to the development of PCC phenotypes. Bioinformatic methods include physical interactions using HI-C and Chia-PET analysis, Transcription Factors (TFs) binding ability, RNA structure modeling, epigenetic, and pathway analysis. This SNP resides within two long RNA genes, LINC01276 and FOXP4-AS1, and is located at ~31 kb upstream of a transcription factor FOXP4. This DNA region, including this SNP, physically interacts with FOXP4-AS1 and FOXP4, implying that this regulatory SNP could alter the normal cellular function of FOXP4-AS1 and FOXP4. Furthermore, rs9367106 is in eQTL with the FOXP4 gene in lung tissue. rs9367106 carrying DNA sequences act as distant enhancers and bind with several transcription factors (TFs) including YY1, PPAR-α, IK-1, GR-α, and AP2αA. The G>C transition extensively modifies the RNA structure that may affect the TF bindings and enhancer functions to alter the interactions and functions of these RNA molecules. This SNP also includes an ALU/SINE sequence and alteration of which by the G>C transition may prevent IFIH1/MDA5 activation, leading to suppression of host innate immune responses. LINC01276 targets the MED20 gene that expresses mostly in brain tissues, associated with sleep disorders and basal ganglia abnormalities similar to some of the symptoms of PCC phenotypes. Taken together, G>C transition of rs9367601 may likely alter the function of all three genes to explain the molecular basis of developing the long-term symptomatic abnormalities in the lungs and brain observed after COVID-19 recovery.

Figures

Full text

1. Introduction

The worldwide infection of the COVID-19 pandemic exceeded 700 million with 7 million fatalities by June 2025 and created enormous socio-economic and health losses. A 4.5% to 85% of the patients who recovered from the disease showed persistent symptoms at least until 3 months to years and are considered to have an acute PCC [1]. Based on WHO and CDC, these symptoms may start from early infection of SARS-CoV-2 to serious illness [2]. However, PCC symptoms are often subjective and hard to distinguish from other related health conditions.

Common PCC symptoms include neuronal abnormalities, including brain fogs, memory loss, depression, sleep disorders, etc., caused by virus-induced serotonin reduction leading to impaired hippocampus and memory function [3]. Respiratory distress is due to the virus-induced inflammation in the lungs. Muscular abnormalities, such as fatigue, are explained by mitochondrial dysfunction in skeletal and smooth muscles [4]. Blood coagulation and thrombiinflammation may be attributed to complement dysregulation, platelet and monocyte aggregation leading to tissue injury [5]. The antibody against the spike protein acts as an “abzyme” that induces inflammation in PCC [6]. Manifestation of PCC may appear as a combination of these symptoms happening from time to time [7].

Severity of the COVID-19 disease differs from individual to population level, involving the host genetic factors that might play a crucial role in developing symptoms and mortalities. A worldwide initiation of host-genetics initiative COVID-19 HGI and others using GWAS identified several gene polymorphisms associated with COVID-19 critical illness and sub-phenotypes [8,9,10,11,12]. They also identified a SNP rs1886814 (A>C, Hg38, chr6: 41,534,945) ~11 kb (11,436 bp) upstream of the promoter of the transcription factor, FOXP4 (hg38 chr6:41,546,381–41,602,384, coding region, 41,565,761–41,598,936) that also passes through an RNA gene, FOXP4-AS1. The worldwide branch of COVID-19 HGI, the Long COVID-19 HGI used meta-analysis of GWAS data from patients of 24 studies comprising 16 countries with PCC (n = 6450) patients and control (n = 1,093,995), those who did not develop PCC symptoms), identified a leading SNP rs9367106 (G>C; Chr6: 41,515,652 bp) ~31 kb (30,829 bp) upstream of FOXP4 promoter and ~19 kb (19,393 bp) upstream of rs1886814 that is associated with PCC [13,14]. We also demonstrated that the association of rs9367106 with PCC is an independent risk factor and cannot be explained by severity alone. Association of rs9367106 with PCC for developing persistent symptoms over COVID-19 critical illness mediated by rs1886814 may require additional involvement of altered FOXP4 function or other genes.

When an SNP is associated with a complex disease, it generally alters multiple cellular pathways to develop the complexities of symptoms of that associated disease [15,16]. Here, I, in Silico, dissected the molecular functions and probable pathways of G>C transition of rs9367106 to dysregulate LINC01276, FOXP4-AS1, and FOXP4 expression. rs9367106 is located within two Long noncoding RNA, LINC01276, FOXP4-AS1 and upstream of FOXP4. Based on its location, rs9367106 is an SNP for regulatory function, and a G>C transition could alter (1) the molecular function of these two RNAs by changing the RNA structure and (2) by inhibiting transcription factors (TFs) binding to dysregulate these two gene expressions, including FOXP4. Two RNA genes are mostly expressed in various brain tissues, including other tissues, whereas FOXP4 is mostly expressed in the lung, brain/hypothalamus, liver, kidney, and intestines. G>C transition indeed modifies the structure of RNAs. Identification of several TFs that bind to this SNP region explains the alteration of molecular pathways. The effect of the G>C transition in altering molecular pathway likely facilitates abnormal brain and lung functions, and inflammation, which are mostly affected in PCC syndrome.

2. Results

2.1. The rs9367106 Resides Within the Long Noncoding RNAs, LINC01276 and AS-FOXP4 and Physically Interacts with the Coding FOXP4 Gene

The SNP rs9367106 (hg38, Chr6: 41,515,652 bp) is located in the ~31 kb upstream of a protein coding transcription factor FOXP4 and within the two long noncoding RNAs, LINC01276 (hg38, 41,502,444 bp-41,519,852 bp, size: 17,409 bp) and FOXP4-AS1 (hg38 chr6: 41,437621–41,546,315 size: 108,695 bp). FOXP4-AS1 is a less-studied RNA gene, and LINC01276 is a long noncoding RNA whose functions are moderately explored.

Both ENCODE data of Hi-C (High-C) experiments in adult lung cells and Mi-C (Micro-C) dataset of h1ESC cells demonstrate that this SNP region resides in TADs (Topologically Associated Domains) that enable it to physically interact with the promoter and coding region of FOXP4 [Figure 1a,b, Table 1]. Further, the analysis of all Hi-C and Mi-C datasets in various cells (h1ESC, HFF6, K562, IMR90, GM12978) [17,18,19,20,21,22,23] indicates the physical interactions of the rs9367106 region with FOXP4 [Table 1, Figure 1 and Figure 2]. In most analyses the contact domain (TADs) spanned from -49 kb upstream of the SNP (Hg19, 41,335,000) to the end of the FOXP4 gene (41,570,122 bp, Hg19), confirming the physical interaction between LINC01276, FOXP4-AS1, rs1886814, and FOXP4. The SNP region also resides within the interacting loop at least in a minimum distance of ~5 KB in a dataset in K562 cells (Hg19, 41,480,511) [23]. However, the interacting positions are not precise, and some of the rs9367106-FOXP4 region appears to be located in both TADs and interacting loops. Further, the SNP rs9367106, FOXP4-AS1, and FOXP4 reside in the same compartment in the nuclear scaffold, strengthening their physical interaction [Supplementary Figure S1].

The physical interactions of these two regions are confirmed by Chia-PET experiments with POL2RA antibody [24,25] in K562 cells [Table 2, Supplementary Figure S2]. The interacting region of rs9367106 is not limited to only the FOXP4 gene and may interact with other genes too. However, the rs9367106 carrying both RNAs physically interact with each other, implying that the G>C transition carrying LINC01276 and FOXP4-AS1 likely regulates the FOXP4 expression.

2.2. The G>C Transition of rs9367106 Extensively Alters the RNA Structure

This SNP is located within the sequences that code for two RNA genes, indicating whether the effect of the G>C transition has an impact on RNA structure. The rs9367106 flanking 100 bp sequences (total 200 bp including both sides), when tested for RNA structure modeling, in both MFE (Minimum Free Energy-base pairing) and centroid (centrafold) models, the transition indeed showed extensive structural differences in sequences carrying mutant (MT) (C) than Wild Type (WT) (G) allele [Figure 3a]. The MFE is calculated by the base pairing of the loop structures and indicates stability of the RNA secondary structure that is lower in WT (MFE, −72.90 kcal/mol (MFE model) and −69.00 kcal/mol (centroid model) than MT (MFE, −52.50 kcal/mol (MFE model) and −39.70 kcal/mol (centroid model) suggesting that WT RNA structures are more thermodynamically stable than MT. The differences in RNA secondary structure may affect the enhancer function and TFs binding, leading to alterations in cellular functions in MT than in WT.

2.3. The rs9367106 Disrupts an ALU/SINE Element

Extensive search of epigenetic databases identified that the rs9367106 carrying sequences encompass an ALU(Y) sequence that acts as a SINE element [Figure 3b]. The boundary of ALU/SINE sequence is 41,483,213–41,483,513 bp, Hg19 (Hg38, 41,515,575–41,515,775 bp). ALU/SINE sequences of the host genome commonly mimic the dsRNA of a virus that is formed as intermediates during viral replication [26]. Viral dsRNA activates innate immunity through MDA5, which acts as a major cellular sensor protein for capturing SARS-CoV-2 dsRNA during infection [27,28,29]. In a human cell culture model, disruption of ALU sequences by ADAR1 (Adenosine Deaminase At RNA) prevents IFIH1/MDA5 activation. Thus, the G>C transition should disrupt the rs9367106 ALU/SINE elements as it changes RNA structure and prevents host immune activation against SARS-CoV-2, that may lead to increased inflammation.

2.4. rs9367106 Carrying DNA Region Binds with Several Transcription Factors (TFs)

The SNP flanking 100 bp binds with numerous TFs [Figure 4a,b]. Most notably, this SNP, including 2–5 bp flanking sequences, binds with several TFs, such as YY1, PPAR-α, IK-1, GR-α, and AP2αA. The transition of G>C is expected to alter these TFs’ binding, leading to perturbing various molecular functions.

2.5. rs9367106 Affects the Expression of FOXP4 and MED20 Genes

The eQTL information of rs9367106 is absent from the gTEX portal eQTL databases. However, the eQTL information was obtained in the QTLbase database, where rs9367106 shows eQTL with FOXP4 in lung tissue (p = 1.6 × 10^−8^) and less significant with TREM1 (p = 3 × 10^−4^) and FRS3 (p = 0.0019) in kidney, and TAF8 (p = 0.0047) in eye.

I enquired whether rs9367106 changes the expression of FOXP4, FOXP4-AS1, and MED20 genes in PCC patients. I re-analyzed the RNA seq data of Ryan et al., consisting of 83 individuals with healthy subjects and PCC patients PBMC [30]. They performed RNA sequencing of a series of mild, moderate, severe, and critical COVID-19 patients, of whom all developed PCC symptoms after recovery from COVID-19 and compared them with healthy subjects, those who never had COVID-19. As the rs9367106 is located within the LINC01276, it would express and present in RNA sequencing data from where the genotype of the patient for rs9367106 could be deduced. However, LINC01276 is very low-expressing RNA, particularly in this SNP region, and in many patients, more than 2 sequence fragments are not represented, from which the correct genotype could not be deduced. Many patients’ RNA sequence data are not informative, and I obtained data from 54 patients’ RNA sequences that have at least 1 sequence carrying rs9367106. Thus, I took the strategy as the presence of at least one C (risk allele) would represent a heterozygous. Three patients have at least one C (risk allele). The expressions of FOXP4, FOXP4-AS1, LINC01276, and MED20 gene expression were obtained by counting the highest expression of an exon in each gene for each patient. The analysis showed that the rs9367106 risk allele (C) indeed increases the FOXP4 gene expression significantly [Figure 4c]. It appears that the risk allele also increases the MED20 gene significantly but appears to have no effect on FOXP4-AS1 and Linc0RNA1276 itself. Notably, the other SNP, rs1886814, which is only associated with COVID-19 severity, insignificantly reduces FOXP4 expression but has no distinct effect on LINC01276, FOXP4-AS1, or MED20 expression.

When the overall expression of LINC01276, FOXP4, FOXP4-AS1, and MED20 genes was assessed from the normalized expression of the same RNA sequencing experiments [30], the FOXP4 and MED20 gene expression was increased in moderate and severe patients who developed PCC phenotypes compared to healthy individuals [Figure 4d]. However, both gene expressions went down in critically affected patients. Most notably, FOXP4 and MED20 gene expressions are synchronized with the level of LINC01276, albeit expressed at a very low level. The marked differences in FOXP4-AS1 expression are not evident in PCC patients compared to healthy individuals.

2.6. The SNP-Carrying RNA Region Acts as a Distant Enhancer

A flanking 200 bp DNA sequence including the SNP rs9367106 acts as an enhancer. Further exploration revealed that this enhancer function is known to act on CD4+, CD8+, and HepG (Hepatocellular carcinoma) cells (SEdb2.0 database, [31]). The dysregulation of CD4+ cells is especially important in developing PCC syndromes [32,33].

The search for the rs9367106 SNP-specific superhancer identified its enhancer functions in the embryo, small intestine, stomach, and colon, but not in lung tissue, and that may be due to inadequate information in the database [Figure 5a]. However, in all tissues, these superenhancers act on FOXP4 and SP2 genes, and also with other associated genes. However, these distant functions of these enhancers correlated with the physically interacting regions of the SNP-carrying region [Figure 5b].

2.7. The LINC01276 Targets the MED20 Gene That Extensively Expresses Tissues in the Brain and Modulates Sleep Disorder

Both Hi-C and Chia-PET physical interaction analysis suggest that rs9367106 carrying LINC01276 directly interacts with MED20 promoter and the coding region [Table 3, Supplementary Figures S1 and S3, Figure 2]. LINC01276 is extensively expressed ubiquitously in many human tissues especially, in brain tissues including cerebral cortex, astrocytes, cerebellum, spinal cords, and moderately expressed in lung, testis, and in other tissues [Supplementary Figure S4]. Search for target of LINC01276 revealed that it exerts its promoter enhancer (GH06J041517) function on six genes, FOXP4, FOXP4-AS1, MED20, TAF8, LOC12267957, ENSG00000124593 (www.genecards.org) [34], While LOC12267957, ENSG00000124593 genes are less studied, MED20 showed relevance with some of the PCC phenotypes concerning neuronal symptoms. MED20 (chr6:41,905,354–41,921,139 (GRCh38) is a part of transcription complex involving RNA polymerase for preinitiation of transcription and is extensively expressed in various brain tissues, along with lungs, small intestines, and testis [Supplementary Figure S5]. Further, MED20 gene mutations are associated with familial infantile basal ganglia degeneration and brain atrophy [35]. GWAS studies identified that MED20 polymorphisms are associated with adult sleep disorders [36]. These phenotypes are similar to some of the neuronal phenotypes of PCC syndrome [1]. The modulation of MED20 expression by LINC01276 could have enormous significance in developing some of the neural phenotypes of PCC symptoms.

2.8. FOXP4-AS1 Function

FOXP4-AS1 (FOXP4-antisense 1) is a long noncoding RNA gene that is implicated in playing a critical role in lung, colon, and osteosarcoma [37,38,39]. FOXP4-AS1 sequesters miR-3184-5p to upregulate the expression of FOXP4 in prostate cancer cells [40]. This gene is expressed mainly in the heart, liver, colon, and kidney [Supplementary Figure S6], but its mechanism of molecular functions is mostly unknown. However, FOXP4-AS1 physically interacts and targets FOXP4 [Table 1 and Table 2]. The altered RNA structure and altered TFs binding similar to LINC01276 might change its interactions with FOXP4 to perturb FOXP4 expression.

2.9. FOXP4 Could Alter the Functions of Lung Alveolar Cells and Brain Tissues Leading to PCC Phenotypes

FOXP4 is a FOX family transcription factor that binds a specific sequence. This gene is mostly expressed in the lungs, brain, kidneys, liver, and testis, etc. [Supplementary Figure S7]. Single-cell RNA seq revealed that FOXP4 expresses differentially and specifically in various types of lung cells, especially in alveolar cells and in various brain cells, including the hypothalamus [Supplementary Figures S8 and S9]. FOXP4 interacts with other FOX-family members and SOX2, MK167, etc., and the FOXP4-FOXP1 axis that is associated with several phenotypes of FOXP1 syndrome [Supplementary Figure S10]. FOXP4 plays a critical role in lung and brain development, and the other FOXP4 mutations are associated with non-small cell lung carcinoma and adenocarcinoma through the FOXP4-SOX2 axis [41,42,43]. Further, the FOXP4 SNPs are associated with neurodevelopmental disease, sensorineural hearing loss, speech/language disorder, and brain development [44,45,46]. Thus, alteration of FOXP4 expression due to the G>C transition in rs9367106 is expected to contribute to some of the lung and brain phenotypes of PCC syndrome.

2.10. rs9367106 Co-Expressed Genes Induce FOXP4-FOXP2 and TP63 Enriched Pathway

Co-expression of genes often indicates their interactions to modulate molecular function. rs9367106 carrying DNA codes for two noncoding RNAs and using the ARCH4 RNA seq data, it created a co-expressed Geneset. The top significantly co-expressed 35 genes [Supplementary Table S1] are used to model the transcription pathway analysis and molecular function [Figure 6a]. FOXP4-FOXP2 and WT (TP63) pathways are the most prominent molecular functions of this Geneset [Figure 6b]. Other notable pathways are PPARG, EPAS1, and ESR2, which are mostly affected by these genes. Enrichment of transcription pathways is shown in a scattered plot [Figure 6c]. The enrichment of most prominent molecular functions is explored, which suggests prominent molecular functions, such as transcription factor binding, cell adhesion, and hydrolase activities [Figure 6d]. The cell/tissue-specific expressions of these gene functions active sites are mostly localized in the small intestine, lungs, large intestine, and various specific brain tissues, such as the mantle zone, intermediate striatum, and hypothalamus [Supplementary Figure S11]. However, the further role of these pathways could be established to dissect the molecular mechanism involving the development of PCC syndrome.

3. Discussion

PCC syndrome has numerous phenotypes that imply abnormalities involving multiple cellular pathways in various human tissues. However, most conspicuous observed phenotypes affect brain, lung functions, muscle, and vascular functions [7]. The Long COVID-19-HGI consortium considered 46 phenotypes of PCC syndrome, and a GWAS study from patients all over the world identified rs9367106 as a causative polymorphism associated with these conditions.

When a regulatory SNP is associated with a complex disease, it is generally observed that the SNP mostly alters multiple molecular functions [15,16]. The location of the SNP within two long RNA molecules, LINC01276, and FOXP4-AS1, and the upstream of the FOXP4 gene implicates that the transition of G>C in this SNP may alter functions of all three genes/proteins [Figure 5c].

The most important observation is that the G>C transition alters the structure of RNA molecules that may confer its effect on LINC01276 and FOXP4-AS1. The structural changes presumably affect the binding of specific TFs at the adjacent SNP sites that perturb the function of these sequences when acting as enhancers in several cell types. Although enhancer functions in the list of various cell types are not extensive, and they are to be further explored. Given the expression of LINC01276 in CD4+ and CD8+ cells, it should have a profound effect as they generate an adaptive response upon SARS-CoV-2 infection [47]. CD4+ T helper cells are broadly important for antibody generation and reduction of T+ cells in COVID-19 that are associated with efficient antibody neutralization [48]. T cells escape in certain COVID-19 patient populations, resulting in loss of epitope-specific responses [49]. The G>C transition in LINC01276 is expected to disrupt the CD4+ and CD8+ T cell-specific functions in COVID-19 pathophysiology.

It is also noted that the G>C transition could also disrupt the ALU/SINE elements. SINE elements act as cellular dsRNA that mimics the viral dsRNA, which forms as an intermediate during replication and activates the intracellular sensor protein IFIH1/MDA5 that captures the viral RNA, especially for SARS-CoV-2 dsRNA [50,51]. Upon SARS-CoV-2 binding, MDA5 activation is necessary for activation of the innate immune system of the host [11,27,28]. Disruption of ALU/SINE elements with ADAR1 inhibits MDA5/IFIH1 activation against viral infection [26,51,52]. IFIH1/MDA5 is the major PRR (Pattern Recognition Receptor) for SARS-CoV-2 to stimulate interferon (IFN) activation and is involved in developing critically ill COVID-19 phenotypes [11,16,29,53,54]. Thus, disruption of SINE elements by G>C transition should facilitate SARS-CoV-2 invasion and inactivate IFN production that is necessary for preventing SARS-CoV-2-mediated development of COVID-19 severe phenotypes. Epigenetic targeting of ADAR1 has been proposed to be effective for certain cancers [55]. Thus, the epigenetic targeting of ADAR1 could also be tested for preventing PCC symptoms.

The eQTL analysis suggests that rs9367106 indeed modulates the expression of FOXP4. Further, the expression analysis indicates that FOXP4 and MED20 expressions are altered in PCC patients compared to healthy individuals, and their expressions are synchronized with LINC01276 expressions. However, these data are from blood tissues RNA sequencing project that limits the ability to elucidate the full spectrum of expression alterations of these genes in other tissues in PCC.

LINC01276 appears to be most appropriate as it targets and modulates the function of the MED20 gene, which is known to play a critical role in various brain tissues and the cerebellum. The MED20 mostly expresses in the neural cortex, astrocytes, spinal cord, and basal ganglia. Several MED20 mutations are identified in basal ganglia syndrome in familial cases, implicating its importance in brain functions and severe sleep disorder in adults [35,36], suggesting that abnormalities in MED20 function may lead to PCC brain phenotypes similar to brain disorders.

FOXP4-AS1 functions are not extensively investigated, except that it is associated with lung, prostate, and liver cancers [37,38,39]. However, in prostate cancer cells, FOXP4-AS1 sequesters miRNA-384-5p to upregulate FOXP4 functions, suggesting that the role of this miRNA could be investigated in developing PCC phenotypes. The expression of this FOXP4-AS1 is mostly limited to the liver, kidney, testis, and other organs, but not extensively in brain tissues. However, the interaction of FOXP4-AS1 with FOXP4 mediated by miRNA-384-5p and their direct physical interactions imply that the transition of G>C may alter the FOXP4 expression, leading to abnormal cellular pathways, which may develop to PCC symptoms.

The rs9367106 SNP region physically interacts with the FOXP4 promoter and within the coding regions. Further, LINC01276 also interacts with FOXP4 as one of the targets and implicates that the G>C transition should alter the function of the FOXP4 gene [34]. FOXP4 is mostly expressed in lungs, liver, and kidneys, and moderately in brain tissues, and its polymorphisms are associated with prostate and liver cancer [56,57].

The enrichment analysis of the Co-expressed Geneset of rs9367106 indicates that it can perturb many genes in the FOXP4-FOXP2 pathway and is involved in critical molecular functions, such as DNA binding activity, cell adhesion, and protein binding activities. However, extensive analysis of the FOXP4 and its role in developing PCC phenotypes could be further explored. Nevertheless, the G>C transition in rs9367106 may alter the functions of LINC01276, FOXP4-AS1, and FOXP4 involving cellular pathways that could explain the development of some of the PCC symptoms [Figure 5c].

Extensive experimental verifications are needed to support the association. The most appalling supposition is the alteration of RNA structure as rs9367106 encompasses two RNAs, but it is not easily plausible due to the large size of LINC01276 (17,409 bp) and FOXP4-AS1 (108,695 bp). Using short sequences including the SNP could be experimentally verifiable in vitro, although it will not directly provide information about the development of PCC phenotypes. However, genome editing the mutation with CRISPR/CAS9 in respiratory or brain cell lines and tracking the LINC01276 functions would result in valuable information about the mechanism of risk allele- specific functions.

Physical and eQTL interactions are important aspects to understand the mechanism of regulating other genes, such as FOXP4 and MED20. Using PCC patient cells and designing HI-C/MI-C experiments would confirm these interactions. Analysis of vast RNA sequence data (with sufficient depth) of an ethnic population (for example, the East Asian population) would confirm the eQTL interactions of this SNP with FOXP4 and MED20.

Enhancer and TFs binding functions could be easily assessed with allele-specific CHIP sequencing and luciferase assay, respectively, but these two methods also cannot shed light on developing PCC phenotypes, although they are useful to confirm the differential effect of the risk allele. Disruption of the ALU/SINE sequence is the most important aspect, as the designs of experiments are feasible with CRISPR/CAS9-edited respiratory cells. Assaying inflammatory markers in these edited cells may explain the mechanism of abrupt inflammation observed in PCC patients.

Limitation

This study is solely based on bioinformatic analysis that relies on relevant genomic databases. As the association of an SNP, especially with a complex disease, does not equate causation, relevant experimental verifications are needed to validate the association. The major limitation is in eQTL analyses where sufficient heterozygous patients were not obtained due to non-ethnicity match (RNA sequencing was performed on a mostly Caucasian population where MAF (minor allele frequency) is 0.4–0.16) and less sequencing depth. LINC01276 itself is a low-expressing RNA, and this SNP region is specifically less represented due to unknown reasons. Further, large sequencing data from the East Asian population is not available with higher sequencing depth. However, further analysis of large RNAseq data with sufficient depth from relevant population or already genotyped individuals would eventually be needed to confirm the eQTL of rs9367106 with the FOXP4 and the MED20 gene. Nevertheless, this study generated numerous future experimental directions that could be tested.

4. Materials and Methods

4.1. Physical Interaction

Hi-C physical interaction maps are derived from the ENCODE [19] lung cancer samples with 5 kb resolution and the MI-C dataset of h1ESC (human embryonic stem cells) from the Decker laboratory [17] with 2 kb resolution. These data are analyzed using Juicebox (version 1.11.08) interacting software [58] (https://github.com/aidenlab/Juicebox/ (accessed on 10 January 2025)).

The interacting domains (TADs, Topologically Associated Domains, loops of interactions, and compartment analysis were also performed in Juicebox. The interacting regions of rs9367106, FOXP4-AS1, and FOXP4 genes are presented in Hg19, which is −32 kb different from Hg38 in this region. Compartment analysis was performed in the MI-C dataset of h1ESC cells by software Eigenvector (accessed on 10 January 2025) (https://github.com/aidenlab/Juicebox/) [58] at 500 kb resolution (the maximum resolution available in this software).

The physical interaction of the rs9367106 carrying region with the FOXP4 gene was also confirmed by Chia-PET [24] with POL2RA antibody in K562 cells. These analyses were performed in the Chia-OET browser at http://3dgenome.fsm.northwestern.edu/index.html (accessed on 16 October 2024) [25]. For physical interaction between the LINC01276 (rs9367106) and the MED20 gene are also deduced and confirmed in Hi-C and Chia-PET using the same browsers.

4.2. eQTL Analysis

SNP-based eQTL analysis of rs9367106 was performed in QTLbase (http://mulinlab.org/qtlbase) (accessed on 26 January 2025) [59].

eQTL analysis in RNA sequencing data of PCC patients was performed from Ryan et al. [30]. The etiologies, patient characteristics, and healthy subjects are given in Table 4. rs9367106 is located within the LINC01276; thus, the RNA sequence carrying variants is identifiable in RNA sequencing data, and the genotype can be deduced from the expression of the variant. For eQTL analysis in PCC patients.

4.3. Expression

RNAseq expression maps for LINC01276, FOXP4-AS1, and MED20 were derived from the RNAseq experiments in various human tissues that are visualized in Genecards (https://www.genecards.org). Single-cell RNA sequencing maps for FOXP4 are obtained from ProteinAtlas (https://www.proteinatlas.org).

4.4. Enhancer Analysis

Approximately 100 bp flanking sequences of each side of rs9367106 (total 200 bp) are used for the enhancer search in enhancer atlas (http://www.enhanceratlas.org/). The SNP-specific superenhancer was analyzed in SEdb 2.0 (https://bio.liclab.net/SEanalysis/ accessed on 6 July 2025)) [31]. The target of each superenhancer carrying the SNP was assessed.

4.5. Transcription Factor Binding Sites

The transcription factor (TFs) binding sites are obtained by searching TFs option along the same 200 bp DNA sequences flanking the SNP in https://alggen.lsi.upc.es/cgi-bin/promo_v3/promo/promoinit.cgi?dirDB=TF_8.3 (accessed on 5 October 2024) [60,61].

4.6. RNA Structure Modeling

RNA structure modeling was performed in RNAFOLD (http://rna.tbi.univie.ac.at//cgi-bin/RNAWebSuite/RNAfold.cgi) (accessed on 9 July 2024) [62,63]. The same 200 bp DNA sequence (used for enhancer search) was transcribed to RNA sequences for modeling with the Wild Type (WT) allele (G) or risk allele (C). The modeling gives output of MFE-base pairing (Minimum Free Energy) and centroid secondary structures. The minimum free energy of maximally stabilized structures was obtained in both cases.

4.7. ALU/SINE Sequence Identification

The same 200 bp flanking rs9367106 genomic sequence was used to search for epigenetic databases in WUSTL (Washington University Epigenome database (https://epigenomegateway.wustl.edu) (accessed on 5 July 2024).

4.8. LINC01276 Target Identification

LINC01276 target genes were identified from the interactome map of Genecards (https://www.genecards.org). The interacting genes are listed, and the MED20 gene is followed up with physical interaction, expression, and functions.

4.9. Co-Expression, Gene and Pathway Enrichment

Co-expression, gene and pathway enrichment are performed by the software, EnrichR (https://maayanlab.cloud/Enrichr/#about) (accessed on 18 January 2025) [64,65,66]. First, a geneset of co-expressed genes is created using the variant rs9367106. Second, this Geneset was used to model the TFs perturbation, tissue-specific expression, and their relevant pathways. The molecular functions of these genes of the Geneset are modeled by the software using Funrich (Version, 3.1.3) [67].