Novel LRF/ZBTB7A variants and known HbF-modulating SNPs in transfusion-dependent β-thalassemia Authors: Yunus Arikan, Tugba Karaman Mercan, Merve Embel, Erdal Kurtoglu Published: 2025 ABSTRACT AbstractBackgroundβ-Thalassemia is a hereditary blood disorder with a highly variable clinical presentation that is partly influenced by genetic modifiers that regulate fetal hemoglobin levels. Elevated HbF can ameliorate the clinical symptoms of β-thalassemia, making the identification of HbF-modifying loci crucial for understanding disease variability and developing potential therapeutic strategies.MethodsIn this study, we evaluated the distribution and effect of known HbF-associated single nucleotide polymorphisms within the BCL11A, HMIP, and XmnI-HBG2 loci in β-thalassemia patients. Our investigation was initiated by a hypothesis based on genomic data from the K562 cell line, which indicated the presence of a specific LRF/ZBTB7A variant. In addition, we explored the role of LRF/ZBTB7A, a transcription factor implicated in globin gene regulation, through Sanger sequencing and in silico analyses. A hypothetical Genetic Modifier Score was devised to quantify the cumulative effect of the five major HbF-associated QTLs.ResultsWhile we confirmed the previously reported associations of the common SNPs with HbF levels, we also identified novel rare variants in LRF/ZBTB7A (p.Pro241Leu, p.Asp344Asp, p.Glu277del) that may influence HbF expression. XmnI-HBG2 accounted for a small yet significant proportion (7%) of HbF variability. A significant positive correlation was found between the Genetic Modifier Score and HbF levels, yet the model explained only ~ 14% of the variance, highlighting a substantial role for other modifiers such as the novel LRF/ZBTB7A variants identified here.ConclusionsThese results suggest LRF/ZBTB7A is a novel modifier of HbF. The discovery of the variants, against a backdrop of modest QTL effects, underscores its potential and warrants further functional investigation.Supplementary InformationThe online version contains supplementary material available at 10.1186/s12920-025-02275-5.Keywords: ZBTB7A/LRF, Fetal hemoglobin, Beta thalassemia, K562, Bioinformatics Background Beta thalassemia (OMIM #613985) is the most common hereditary blood disease and is a Mendelian disease with a well-known genotype‒phenotype correlation and autosomal recessive inheritance [1]. Homozygous or compound heterozygous mutations in the beta globin (HBB) gene (*141900) cause beta thalassemia. The amount of fetal hemoglobin (α2γ2) is among the genetic factors that affect the clinical outcome of this disease [2]. Increasing the amount of fetal hemoglobin is one of the strategies that has been implemented in beta-hemoglobinopathies [3, 4]. Fetal hemoglobin induction via lentiviral gene therapy and CRISPR-based clinical trials (NCT03655678, NCT03745287, and NCT05477563) are promising for the treatment of beta thalassemia or sickle cell disease [5, 6]. The modifier genes that endogenously determine the fetal hemoglobin level (quantitative loci for the hereditary persistence of fetal hemoglobin (QTL)) are the HBG2-XmnI (rs7482144) polymorphism, the intergenic variants in the HBS1L-MYB (HMIP) locus, and the single nucleotide polymorphisms in the BCL11A (*142335) gene [7]. The LRF/ZBTB7A (*605878) gene, which is impaired in the pathogenesis of acute myeloid leukemia, was found to be an independent repressor of the fetal hemoglobin induction pathway at the same time [8, 9]. Despite advances in understanding HbF regulation, interindividual variation in response to known modifiers suggests the presence of additional genetic factors. In the present study, we genotyped 2 SNPs in the BCL11A gene together with 2 SNPs in the HMIP region and the XmnI-HBG2 polymorphism via the reverse dot blot hybridization-based method (RDBH) in 58 patients with TDBT. We first showed how much the 5 most frequently studied SNPs could increase the amount of HbF in our beta thalassemia cohort. Later, as a result of bioinformatics screening for the detection of specific candidate variation (LRF/ZBTB7A, p.P241L), we revealed novel mutations of a candidate gene that could affect only the amount of HbF for the first time in our cohort. Methods Study design and ethical approval The local ethical committee approved the current prospective study (KAEK-189_2019.03.13_12). Written informed consent was given by all participants in accordance with the Declaration of Helsinki. DNA isolation (QIAAmp® DNA extraction blood minikit, Qiagen) was performed from the peripheral blood of 58 TDBT patients and 60 controls who signed informed consent forms. The HbF levels of patients (HbF > 2%) and controls (HbF ≤ 2) were determined via the Bio-Rad Variant™ Hemoglobin Testing System II. Genotyping strategy Comprehensive HBB genotyping to confirm the beta-thalassemia diagnosis was performed only for the TDBT patients, using data retrieved from their clinical archives. For the control group, HBB genotyping was not clinically indicated; their inclusion served to establish a population baseline for LRF/ZBTB7A variant screening and HbF distribution. Genotyping for 5 SNPs (rs7482144, rs1427407, rs10189857, rs28384513, and rs9399137) was performed via the RDBH-based strip method, ViennaLab StripAssays® (ViennaLab Diagnostics GmbH). This assay is based on the use of subsequent multiplex PCR and RDBH. To best reflect endogenous production, HbF levels were measured from blood samples obtained just before a scheduled transfusion. Targeted Sanger sequencing based on cell line hypothesis Our investigation was initiated by a hypothesis based on prior genomic data from the K562 cell line, which indicated the presence of the specific LRF/ZBTB7A p.P241L (c.722 C > T) variant. To test the clinical relevance of this observation, we designed a targeted Sanger sequencing approach. Genomic DNA from 58 TDBT patients and 60 controls was used to amplify the specific region of exon 2 of the LRF/ZBTB7A gene (NC_000019.9) harboring the c.722C > T locus using the following primers: forward, 5’-ATGACCTGGATGCCACCAAG-3’; reverse, 5’-GCCTTGGCTCGGATCTTCTT-3’. The resulting PCR products were sequenced on an ABI-3130-XL Genetic Analyzer (Applied Biosystems). The possible damaging effects of the identified mutations were evaluated by considering the American College of Medical Genetics and Genomics (ACMG) guidelines and Mutation Taster database. In silico analyses for splice prediction (SpliceAI), evolutionary conservation (PhyloP), codon usage (https://www.bioinformatics.org/sms2/codon_usage.html), and pathogenicity (CADD, AlphaMissense) were also performed. Variant-specific ACMG pathogenicity criteria, gnomAD allele frequencies, and population-specific allele frequencies for Turkiye for the ZBTB7A gene variations were individually retrieved from the https://search.ngscloud.com/portal. Data analysis Nonparametric Mann‒Whitney U and Kruskal‒Wallis tests were performed via SPSS version 20 (IBM SPSS Statistics for Windows-MacOS) to assess the effects of relevant SNPs on HbF levels. The extent to which SNPs could explain the HbF level was determined via univariate analysis. Linkage disequilibrium (LD) scores of two SNPs of each BCL11A and HBS1L-MYB were retrieved from the database for all human populations (https://ldlink.nih.gov/?tab=ldpop). The genotypic and allelic frequencies of the SNPs in our patient group were also calculated. Genetic modfifier score calculation To quantitatively assess the cumulative contribution of known Hb F-associated quantitative trait loci (QTLs), a hypothetical Genetic Modifier Score was devised. The score was based on five major QTLs (specify loci if desired, e.g., BCL11A, HBS1L-MYB, and XmnI-HBG2). For each SNP, genotypes were assigned a value of 0, 1, or 2 points based on established literature, corresponding to the number of HbF-increasing alleles. The individual SNP scores were then summed to create a cumulative Genetic Modifier Score for each patient, where a higher score indicates a greater predicted genetic predisposition for elevated HbF production. A patient with a genetic modifier score of “2-1-0-0-0” would receive a total of 3 points (calculated as 2 + 1 + 0 + 0 + 0). The theoretical maximum of the genetic risk score is 10 points. Results The study followed a three-step design, including QTL genotyping, candidate mutation selection in LRF/ZBTB7A, and screening in both patients and controls, as shown in Fig. 1. The mean age of the patients was 32.4 ± 9.3 years, and the mean HbF level was 8.8 ± 5.6%, with values ranging from 2.0% to 24.4%. The presence of the T allele of XmnI-HBG2 was significantly associated with higher HbF levels (p = 0.032, Mann‒Whitney U test), even though the overall genotype effect on HbF was not statistically significant (p > 0.05, Kruskal‒Wallis tests). There was also a significant difference between being above the mean HbF value of 8.8 and having the T allele (p = 0.046, Fisher’s exact test). Furthermore, we showed that the XmnI-HBG2 polymorphism explained 7.0% of the variance in HbF levels after adjusting for age (ANCOVA, F(2,54) = 4.093, p = 0.048; n = 57 due to one missing age value), marking the first such report in the Turkiye population (Supplementary Table S1). Graphical abstract summarizing the study design and concept. First, known QTLs that cause high HbF levels were genotyped in only TDBT patients via the RDBH assay. Second, regarding the literature and obtaining a deep genomic profile of the K562 cell line, a certain mutation in the exon 2 region of LRF/ZBTB7A was selected, and this mutation was screened in both patients (n=58) and controls (n=60) via Sanger sequencing. A favorable genetic modifier score was calculated for each patient to predict disease severity and clinical outcomes No relationship was found between the presence of a certain allele of the other 4 SNPs and the HbF level. In terms of genotype frequencies (Fig. 2), the SNP with the highest heterozygosity was BCL11A rs1427407 (77.6%), whereas the lowest was XmnI-HBG2 rs7482144 (10.3%). The LD scores of the SNPs (rs1427407 and rs10189857) in BCL11A were 0.25, and the LD scores were 0.13 for the SNPs (rs28384513 and rs9399137) in HBS1L-MYB (Table 1). When the minor allele frequency of SNPs in the normal population was compared with the allele frequencies in our patient population, the greatest difference was detected in rs9399137 in HBS1L-MYB. The relationship between the cumulative genetic modifier score and HbF levels. The BCL11A rs1427407 GT genotype presented the highest heterozygosity rate among the key HbF modulating loci in the TDBT cohort. On the right panel, the score calculated from five major HbF associated loci, demonstrated a significant positive correlation with HbF%, underscoring the collective contribution of these genetic modifiers to fetal hemoglobin variation in the cohort Table 1Baseline characteristics and frequencies of HbF-modifying variantsModifierGenotype (%)%HbF (median)P value*MAFaAllele Freq.bLD (r2)XmnI-HBG2 rs7482144CC (86.2)C allele (7.5)0.740.91CT (10.3)T allele (10.6)0.0320.260.09TT (3.5)BCL11A rs1427407GG (20.7)G allele (7.5)0.800.54GT (77.6)T allele (6.8)0.3280.200.46TT (1.7)0.25 BCL11A rs10189857AA (43.1)A allele (8.1)0.490.58AG (29.3)G allele (7.2)0.1950.510.42GG (27.6) HBS1L-MYB rs28384513AA (58.6)A allele (8.1)0.740.78AC (37.9)C allele (8.4)0.4970.260.22CC (3.4)0.13 HBS1L-MYB rs9399137TT (56.9)T allele (8.8)0.850.28TC (29.3)C allele (8.1)0.1870.150.72CC (13.8)LD Linkage disequilibrium score (r²) retrieved for all populations from the LDlink database*Mann-Whitney U testaThe Minor Allele Frequency reported for the general population in the 1000 Genomes ProjectbThe frequency of each allele calculated from the genotype data within our study cohort Baseline characteristics and frequencies of HbF-modifying variants BCL11A rs10189857 HBS1L-MYB rs28384513 HBS1L-MYB rs9399137 LD Linkage disequilibrium score (r²) retrieved for all populations from the LDlink database *Mann-Whitney U test aThe Minor Allele Frequency reported for the general population in the 1000 Genomes Project bThe frequency of each allele calculated from the genotype data within our study cohort In the LRF/ZBTB7A gene, the investigation was motivated by an initial screening of an archived version of the DepMap database, which contained an entry for the p.P241L variant in the K562 cell line. To assess the clinical relevance of this observation, we performed targeted Sanger sequencing in our cohort. This analysis confirmed the presence of the p.P241L variant (rs538445777, c.722 C > T) in a patient with an HbF value of 3.5% (Fig. 3a). Additionally, a novel in-frame deletion (c.829_831del, p.E277del, rs1361497779) was detected in a control group individual with an HbF value of 1.5% (Fig. 3b), and a synonymous mutation (rs112339723, p.D344D) was found in one patient with an HbF value of 5.1% (Fig. 3c). The rs538445777, rs1361497779 and rs112339723 variations in the LRF/ZBTB7A gene were evaluated as VUSs according to the ACMG criteria (Table 2). Representative Sanger sequencing chromatograms of LRF/ZBTB7A variants observed in the study. a Heterozygous rs538445777 (c.722C>T) in a patient with elevated HbF (3.5%); b) In-frame deletion rs1361497779 (c.829_831delGAG) in a control subject (HbF: 1.5%); c) rs112339723 (c.1032C>T) in a patient with an HbF level of 5.1% Table 2Genetic and hematological profiles of cases with ZBTB7A variantsVariationProtein% MAF (gnomAD)% in TurkiyeACMG criteriaaSNP profile% HbF%HbA2HGB(g/dL)MCV(fL)HBBbrs538445777p.P241L0.050.19PM11-1-0-0-03.52.89.585.4−30 T > A/IVS.I.110 (G > A)rs1361497779p.E277del - - PM2-1.5----rs112339723p.D344D0.140.31BS21–2-0-0-05.12.48.277.9IVS.I.110 (G > A)/IVS.I.110 (G > A)0:homozygous normal, 1:heterozygous, 2 homozygous mutated allele. First two SNPs are located in BCL11A gene, following two SNPs are located in HBS1L-MYB locus. The last SNP (rs7482144) is also known as XmnI-HBG2aThe five HbF modifier SNPs are rs1427407-rs10189857-rs2838451-rs9399137-rs7482144, respectivelybThe complete dataset for all study participants is available in Supplementary Table S2. p.P241L and p.D344D (TDBT patients); p.E277del (control) Genetic and hematological profiles of cases with ZBTB7A variants 0:homozygous normal, 1:heterozygous, 2 homozygous mutated allele. First two SNPs are located in BCL11A gene, following two SNPs are located in HBS1L-MYB locus. The last SNP (rs7482144) is also known as XmnI-HBG2 aThe five HbF modifier SNPs are rs1427407-rs10189857-rs2838451-rs9399137-rs7482144, respectively bThe complete dataset for all study participants is available in Supplementary Table S2. p.P241L and p.D344D (TDBT patients); p.E277del (control) In the archive records (Supplementary Table S2), the HBB gene mutation in the patient whose HbF value was 3.5 was − 30 T > A/IVS.I.110 (G > A). The patient was splenectomized and received a blood transfusion once a month. This patient has the p.P241L mutation in the LRF/ZBTB7A gene, and the HbF modifier SNP profile is 1-1-0-0-0. Although the HBB profile of the individual with an HbF value of 1.5 in the control group is unknown, the only difference in the HbF modifier profile is the homozygous mutation of the rs10189857 SNP in BCL11A (Table 2). The patient was splenectomized and had a homozygous IVSI-110 (G > A) mutation in the HBB gene with a recorded HbF level of 5.1%. To quantitatively assess the cumulative effect of known HbF associated variants (Supplementary Table S3), we calculated a Genetic Modifier Score based on five major QTLs (Fig. 2). This analysis revealed a significant positive correlation between the score and HbF levels (Spearman r = 0.395, p = 0.002), although the linear regression model explained only ∼14% of the variance (R² = 0.14, p = 0.004). Discussion Studies to increase gamma globin expression pharmacologically or by using various genome editing technologies continue to be a focus of current research [10–12]. In addition, various endogenous genomic variations found in individuals are associated with naturally high HbF. In this study, we aimed to investigate the associations between elevated HbF levels and other HbF-related variants in the BCL11A, HMIP, and HBG2 loci as well as potential functional variants in the recently described modifier gene LRF/ZBTB7A. The genetic variants in these three loci account for approximately 20–50% of the variation in HbF levels in patients with SCA and β-thalassemia, as well as in healthy adults [13]. Our univariate analysis suggested that the XmnI-HBG2 polymorphism was associated with an increase in HbF level by approximately 7.0% in our cohort. Similarly, we showed that individuals with the T allele for XmnI-HBG2 had higher HbF levels than the mean HbF level of 8.8 in the patients in our study. One study revealed that XmnI-HBG2 contributes 10.2% of the HbF variance in the healthy European population, whereas SNPs in the BCL11A and HMIP loci are responsible for approximately 15.1% and 19.4%, respectively [14]. A study showing that the XmnI-HBG2 polymorphism, but not SNPs of the BCL11A or HMIP loci, has HbF-increasing properties in the hematologically normal Iranian population [15], similar to our study, although we included patients with high HbF levels in our Turkiye cohort. One study included 100 patients with β-thalassemia and 100 patients with SCA; XmnI-HBG2 polymorphisms, BCL11A (rs1427407) and HMIP (rs9399137, rs66650271), have been associated with both high HbF values and low disease severity [16]. LRF/ZBTB7A was identified as a second independent repressor of HbF (Fig S1), and subsequent studies quickly established the link between KLF1 mutations and elevated HbF levels through the regulation of both BCL11A and LRF/ZBTB7A [17]. Mutations in the LRF/ZBTB7A gene are responsible for macrocephaly, neurodevelopmental delay, lymphoid hyperplasia, and persistent fetal hemoglobin [18] (MNDLFH, OMIM #619769). The patient presented an HbF level of 100% at 1 month of age (reference HbF level at this age: ~70%). At 6 years and 5 months, coincident with the diagnosis of MNDLFH, genetic analysis revealed a de novo missense variant in the LRF/ZBTB7A gene (p.Cys384Trp). In another study in which patients with MNDLFH were involved, the fraction of HbF was found to be elevated in four of the five individuals with LRF/ZBTB7A gene mutations [19]. The heterozygous variants detected in the cohort included p.Glu278* (associated HbF level: 4.4%), p.Thr405Lys (5.3%), p.Val417GlyfsTer123 (11.1%), and p.Arg530GlyfsTer27 (2.2%). We detected a p.P241L change (Fig. 3a) in a patient with an HbF level of 3.5 (Table 2). Moreover, we detected a synonym variation (p.Asp344Asp, Fig. 3c) in a patient with high HbF (5.1%). However, we did not find these two variations in the control group (n = 60), but we detected a novel in-frame deletion (p.Glu277del, Fig. 3b) in the LRF/ZBTB7A gene in an adult female in our control group whose HbF level was 1.5%. Therefore, our study reports the presence of LRF/ZBTB7A gene variants in both patients with TDBT and in individuals from the control group (% HbF level ≤ 2). The allele frequency of c.722 C > T (p.Pro241Leu) was found to be approximately 4 times greater in the Turkiye population than in the gnomAD (0.19% vs. 0.05%) database. With respect to c.1032 C > T (p.Asp344=), the allele frequency in Turkiye is more than 2 times higher than that in gnomAD (0.31% vs. 0.14%). The synonymous variant p.Asp344Asp (rs112339723) was identified in a patient with high HbF (5.1%). In silico analyses (CADD score: 3.14; AlphaMissense: 0.0 ‘likely benign’) did not provide evidence for a deleterious impact or an effect on splicing or local translation efficiency. Therefore, while this correlative finding in a patient with notably high HbF is interesting, the functional relevance of this specific variant remains uncertain. Its effect may be contingent on the genetic context, as this individual was also homozygous for the known HbF-booster SNP rs10189857 in BCL11A. This case highlights the potential for combined effects of known QTLs with other rare or neutral-appearing variants, the significance of which may only become apparent in specific genetic backgrounds and warrant further investigation in larger cohorts. Moreover, the novel LRF/ZBTB7A in-frame deletion (rs1361497779) was identified in a healthy control individual with normal hematological parameters and low HbF (≤ 2%), suggesting its effect on HbF modulation may be context-dependent. LRF/ZBTB7A gene expression levels in patients who received hydroxyurea treatment (homozygous SCA or combined heterozygous sickle cell/beta thalassemia) were found to be lower than those in patients who responded to treatment, as expected [20]. Notably, analysis of the K562 genomic data revealed no evidence of large deletional HPFH mutations, suggesting that its high HbF phenotype may involve alternative regulatory mechanisms such as the LRF/ZBTB7A variants investigated here. Persistent and high fetal hemoglobin levels were achieved in pediatric TDBT patients in whom BCL11A was disrupted by CRISPR-Cas9-mediated genome editing [21]. BCL11A is a promising candidate for treating β-hemoglobinopathies because of its high degree of HbF resurgence, lack of off-target effects, and unaltered gene expression, whereas the HBG1/2 approach also achieves clinically relevant HbF levels with a mediocre safety profile, suggesting that further research could lead to the development of viable alternative gene therapies for such diseases [22, 23]. Our findings are in line with one of the latest CRISPR/Cas9-based studies showing that disruption of either the LRF/ZBTB7A or BCL11A binding site comparably reactivates γ-globin expression, suggesting that both are viable therapeutic targets for β⁰-thalassemia [24]. This is further supported by recent pharmacological evidence demonstrating that targeted degradation of LRF/ZBTB7A robustly induces HbF without significant toxicity, reinforcing its therapeutic potential [25]. The significant yet modest explanatory power (R² ~14%) of the Genetic Modifier Score confirms the aggregate contribution of known QTLs to HbF variation, while also indicating that the majority of the phenotypic variance is driven by other factors. This unexplained variance provides a compelling genetic context for the putative role of novel modifiers, such as the LRF/ZBTB7A variants identified here, which may exert effects powerful enough to override the predisposition set by known QTLs. The statistical power of our study was constrained by the cohort size, which may have been insufficient to detect associations for the other SNPs in BCL11A and HBS1L-MYB, especially given their previously reported modest effect sizes. Furthermore, it should be noted that our study was not designed to compare the modifier score between patients and controls, as the score itself is predicated on the presence of an elevated HbF phenotype. This limitation should be considered when interpreting the non-significant findings for these loci. Conclusions In our study, the XmnI-HBG2 polymorphism accounted for approximately 7% of the variance in fetal hemoglobin levels in transfusion-dependent beta thalassemia patients with high HbF levels. By examining the genomic and transcriptomic data of the K562 cell line, we identified certain variants in the LRF/ZBTB7A gene that we hypothesize may contribute to the variation in HbF levels. Our study demonstrates the utility of leveraging genomic data from model cell lines to generate testable hypotheses for clinical genetic modifiers. Nevertheless, the potential of cell lines to provide researchers with data for a variety of quantitative measurements must be considered and taken into account. Bioinformatics data from cell cultures can reveal pinpoint information about specific characteristics in disease groups that have not been previously investigated. Our study reports novel, rare variants in the LRF/ZBTB7A gene, suggesting their potential as candidate genetic modifiers that may contribute to the variation in HbF levels among TDBT patients. While the correlative data are suggestive, definitive confirmation of their functional impact and therapeutic relevance awaits future studies using targeted in vitro models. Collectively, the modest effect of the cumulative Genetic Modifier Score (R²~14%) and the identification of novel LRF/ZBTB7A variants suggest that a significant portion of HbF variability in our cohort arises from yet to be discovered genetic factors, with LRF/ZBTB7A emerging as a prime candidate for such a modifier. The findings of this study should be interpreted within its limitations. The sample size, while valuable for a rare disease, limits the statistical power for robust subgroup analyses and broad generalizations. Furthermore, the genetic modifier score applied here is a hypothesis-generating model based on literature-derived SNPs, and its predictive value requires validation in larger, independent cohorts. It is also important to note that comprehensive clinical data (such as splenectomy status and HBB genotypes for the entire cohort) was not the primary focus of this genetic association study and was therefore only available and reported for the illustrative cases where a specific LRF/ZBTB7A mutation was found. Finally, the focus on specific SNPs and the lack of functional validation in experimental models mean that other significant genetic factors and the precise biological impact of the identified LRF/ZBTB7A variants may not be fully captured. Supplementary Information Supplementary Material 1. Supplementary Material 2. Supplementary Material 3. Supplementary Material 4.Figure S1. TPM-level expression of LRF/ZBTB7A, BCL11A, and XmnI-HBG2 in K562 cells. K562 cells expressedLRF/ZBTB7A (2.8 ± 0.7 TPM), BCL11A (1.6 ± 1.6 TPM), and HBG2 (0.7 ± 1.6 TPM) transcripts. ACTB (β-actin) is shown as a housekeeping gene for reference. Gene expression values (TPM: transcripts per kilobase million, mean ± SD) were retrieved from the DepMap portal (DepMap ID: ACH-000551; file: OmicsExpressionProteinCodingGenesTPMLogp2.csv) via https://depmap.org/portal/. The right panel shows that LRF/ZBTB7A and BCL11A independently repress XmnI-HBG2 expression. Acknowledgements We thank the patients, their families, and our colleagues for their support. This study was designed to assess HbF modifiers independent of HBB genotype (provided in the supplement). Abbreviations American College of Medical Genetics and Genomics B-cell lymphoma/leukemia 11A (gene) Clustered Regularly Interspaced Short Palindromic Repeats Fetal hemoglobin Hemoglobin subunit beta (gene) Hemoglobin subunit gamma 2 (gene) HBS1L-MYB intergenic polymorphism (locus) Linkage disequilibrium Leukemia/lymphoma-related factor (same as ZBTB7A) Macrocephaly, neurodevelopmental delay, lymphoid hyperplasia, and persistent fetal hemoglobin (syndrome) Online Mendelian Inheritance in Man (database) Quantitative trait locus/loci Reverse dot blot hybridization Transfusion-dependent beta thalassemia Variant of uncertain significance A specific polymorphism (rs7482144) in the HBG2 gene Zinc Finger and BTB Domain Containing 7A (gene) Authors’ contributions Conceptualization: Y.A.; Data curation: Y.A., T.K.M., M.E., E.K.; Funding acquisition: Y.A.; Investigation: Y.A., T.K.M., M.E., E.K.; Methodology: Y.A., T.K.M., M.E.; Project administration: Y.A.; Supervision: Y.A., E.K.; Visualization: Y.A., T.K.M., M.E.; Writing -original draft: Y.A.; Writing&editing: Y.A., T.K.M., M.E., E.K. Funding This study was supported by the Yozgat Bozok University Scientific Research Projects Coordination Unit (Project No: 6602c-TF/19–281). Data availability All data generated or analysed during this study are included in this published article and its supplementary information files. The variant data generated during this study have been deposited in the ClinVar repository under the temporary submission handles SUB15634423, SUB15634458 and SUB15634616.The permanent accession numbers (SCV) for these variants are available and can be accessed via our submitter page on the NIH ClinVar database (https://www.ncbi.nlm.nih.gov/clinvar/submitters/510189). Declarations Ethics approval and consent to participate The local ethical committee (Yozgat Bozok University, Clinical Research Ethical Committee) approved the current prospective study (protocol number: KAEK-189_2019.03.13_12). Written informed consent was given by all participants in accordance with the Declaration of Helsinki. Consent for publication Written informed consent was obtained from all participants for the publication of the data included in this study. Competing interests The authors declare no competing interests. Conflict of interest The authors declare that they have no competing interests. Footnotes Publisher’s Note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. References Associated Data Supplementary Materials Supplementary Material 1. Supplementary Material 2. Supplementary Material 3. Supplementary Material 4.Figure S1. TPM-level expression of LRF/ZBTB7A, BCL11A, and XmnI-HBG2 in K562 cells. K562 cells expressedLRF/ZBTB7A (2.8 ± 0.7 TPM), BCL11A (1.6 ± 1.6 TPM), and HBG2 (0.7 ± 1.6 TPM) transcripts. ACTB (β-actin) is shown as a housekeeping gene for reference. Gene expression values (TPM: transcripts per kilobase million, mean ± SD) were retrieved from the DepMap portal (DepMap ID: ACH-000551; file: OmicsExpressionProteinCodingGenesTPMLogp2.csv) via https://depmap.org/portal/. The right panel shows that LRF/ZBTB7A and BCL11A independently repress XmnI-HBG2 expression. Data Availability Statement All data generated or analysed during this study are included in this published article and its supplementary information files. The variant data generated during this study have been deposited in the ClinVar repository under the temporary submission handles SUB15634423, SUB15634458 and SUB15634616.The permanent accession numbers (SCV) for these variants are available and can be accessed via our submitter page on the NIH ClinVar database (https://www.ncbi.nlm.nih.gov/clinvar/submitters/510189).