The numbers on a routine blood test are strongly heritable, and one of them turned into a medicine: BCL11A was found because it moved fetal haemoglobin, and silencing it now treats sickle cell disease. The same data also warns that several of these numbers mark disease risk without causing it.
Solid lines are connections this site curates. Dashed lines mean the two ends share a research paper — worth knowing, and not a claim that one explains the other.
A full blood count is the most ordered test in medicine. It reports how many red cells, white cells and platelets are in a sample, and how big they are. Most people have seen one. Almost nobody has been told that those numbers are among the most heritable things a laboratory measures.
A 2016 study tested 29.5 million genetic variants against 36 red cell, white cell and platelet properties in 173,480 people. A 2011 meta-analysis in up to 66,867 people found 68 loci for platelet count and volume alone. A 2012 study of up to 135,367 people found 75 loci for red cell traits.
And here is the honest size of it: those 75 loci together explain 4-9% of the variation per trait. Heritable is not the same as explained, and a blood test gives you the actual number in an afternoon.
This is the part worth reading even if you never look at the list below.
Some adults keep making a little fetal haemoglobin — the form of haemoglobin a baby uses before birth — instead of switching entirely to the adult form. How much varies between people, and it is inherited. On its own that is a curiosity: a number on a blood test.
In 2008 genetic association studies traced that variation to a gene called BCL11A, and noticed something else at the same time. People with sickle cell disease and beta-thalassaemia who happened to keep more fetal haemoglobin had milder illness. The same paper that tied BCL11A and the HBS1L-MYB region to fetal haemoglobin levels also tied them to how often people with sickle cell disease had pain crises.
Later that year BCL11A turned out to be the switch itself: the repressor that shuts fetal haemoglobin off during development. In 2013 the common variation found by those association studies was localised to a noncoding enhancer of BCL11A that works only in red cell precursors — which is exactly the piece you would want to disable, because switching the gene off everywhere would do other damage.
Silencing BCL11A in a person's own blood stem cells raises fetal haemoglobin, and fetal haemoglobin does not sickle. That is now a treatment for sickle cell disease.
This site holds rs11886868 in BCL11A and rs7775698 in HBS1L. They are on this page because they move a blood index. What they led to is the best answer this catalogue has to the question of why anyone would study a routine laboratory number.
A 2020 trans-ethnic analysis tested 746,667 people, including 184,535 of non-European ancestry, across 15 blood-cell traits — the first source on this page with real representation outside European-ancestry cohorts. It found 71 associations that European-only analyses miss entirely, plus 28 more found only within a single non-European population. One of them: a missense variant in IL7, found in South Asian participants, linked to both lymphocyte count on a blood test and IL-7 secretion measured directly in the lab — a rare case on this page where a genetic association and its underlying biology were confirmed in the same study.
Red cell distribution width (RDW) — how much red blood cells vary in size within one sample — rises with age and predicts mortality, coronary heart disease and cancer, independent of whether someone is anaemic. A 2017 study of 116,666 UK Biobank volunteers found genetics explains a real share of that variation (29% overall, rising to 33.8% in people over 60 versus 28.4% under 50 — heritability that itself grows with age). It found 194 independent genetic signals, and 71 of them were already known for something else entirely: autoimmune disease, several cancers, BMI, Alzheimer's disease, longevity, age at menopause, bone density, myositis, Parkinson's disease, and age-related macular degeneration. Pathway analysis pointed at telomere maintenance, ribosomal RNA and apoptosis — general aging biology, not a blood-specific process that happens to also predict age.
Positions joined since this page was written
What this is The text above discusses the variants this page was written around. Since then the catalogue has joined 1612 more positions to it, by shared trait or shared paper. They are listed here by the paper each came from; the text does not describe them, and each variant page carries that study's own record.
Chen MH et al. 2020, Cell rs66858280 (TRIM58), rs62510269 (near PKIA-AS1), rs9590569 (ELF1), rs7200918 (SLC7A5), rs1772143 (PM20D1), rs12600858 (LINC01476), rs11646287 (DNASE1), rs71375338 (near MAPK8IP1P1), rs6883116 (MGAT4B), rs4481150 (ITIH3), rs6556313 (RGS14), rs17015169 (TRAF3IP3) and 621 more — PMID:32888493
Vuckovic D et al. 2020, Cell rs55691975 (AC040160.1), rs11023922 (SOX6), rs76445378 (KALRN), rs11741775 (CCDC125), rs61974485 (ARID4A), rs2646420 (RP11-543F8.2), rs3740405 (PPRC1), rs77074394 (KIAA0922), rs148912436 (SPTA1), rs2069443 (CDK5), rs547353809 (SV2C), rs9852648 (PLCL2) and 431 more — PMID:32888494
Sakaue S et al. 2021, Nature genetics rs4667283 (near ITGA4), rs59252007 (HOOK2), rs11170516 (near SP7), rs895692 (near NAT9), rs6904506 (near LINC03066), rs149262078 (ZSWIM6), rs2512412 (UBR5), rs11761603 (near UNCX), rs1123339 (KAT7), rs710177 (YDJC), rs2208590 (PREX1), rs11618989 (near IRS2) and 189 more — PMID:34594039
Jee YH et al. 2025, Nature communications rs7798256 (near RPL15P11), rs9591137 (near PDS5B), rs12535226 (EGFR), rs45563932 (TMEM273), rs34773350 (TFCP2L1), rs12485511 (PLCL2), rs2834320 (LINC00649), rs3741411 (CIMAP1A), rs17116149 (near STN1), rs1570360 (near VEGFA), rs35022307 (SLC2A1), rs1924930 (EDNRB-AS1) and 99 more — PMID:40436827
Kachuri L et al. 2021, American journal of human genetics rs12202615 (near ADGRF2P), rs2665397 (VMP1), rs7098710 (APBB1IP), rs1009181 (H2BC5), rs9402633 (near LINC03002), rs28679728 (near KLHL8), rs4766540 (CUX2), rs16939945 (IRF8), rs2777895 (VMP1), rs8112710 (MARK4), rs86715 (near RN7SL26P), rs9267951 (near TSBP1-AS1) and 73 more — PMID:34469753
Oskarsson GR et al. 2020, Communications biology rs33932084 (PGBD1), rs200484 (HIST1H2BL), rs61742093 (OR2B2), rs4823308 (FBLN1), rs447735 (CDK10), rs73594554 (intergenicN1L), rs117310669 (AKR1C4), rs6570048 (PDE7B), rs10885 (PRRC2A), rs2071466 (TAP2), rs1182933 (HNF1A), rs4073770 (PPARG) and 14 more — PMID:32327693
Timoteo VJ et al. 2023, The Journal of nutritional biochemistry rs199533 (NSF), rs5026246 (near PLEKHM1), rs1265565 (CUX2), rs72826361 (CMAHP), rs4378452 (CUX2), rs994379 (near H3C9P), rs8070723 (MAPT), rs3117574 (MSH5-SAPCD1), rs2498399 (near ZNF322), rs11168351 (near COL2A1), rs2942164 (LINC02210), rs2696671 (near MAPK8IP1P1) and 11 more — PMID:35964923
Astle WJ et al. 2016, Cell rs9267098 (HCP5), rs185523698 (NELFCD), rs10821556 (RP11-349K21.1), rs8064299 (SLC9A3R1), rs9820353 (RPL6P9), rs2507971 (MICB), rs68191221 (KLF3), rs9410380 (near MIR4289), rs185898 (RTTN), rs73076965 (RP4-576H24.2), rs11929092 (LINC00885), rs115487693 (C6orf48) and 10 more — PMID:27863252
Jacobs BM et al. 2024, Nature communications rs907092 (IKZF3), rs377134808 (RAB11FIP3), rs4910566 (HBE1), rs12131809 (MCL1), rs115268031 (CBFA2T3), rs11189159 (near RRP12), rs11128810 (PLCL2), rs73404531 (HBG2), rs72832602 (TRIM38), rs7934446 (near OR51A9P), rs7744284 (CCDC162P), rs55910622 (near SLFN14) and 9 more — PMID:39414775
Hu Y et al. 2021, BMC genomics rs16848876 (MGST3), rs835574 (NOTCH2), rs12122920 (SNX27), rs569025 (LMNA), rs2603961 (NOTCH2), rs520757 (near CDC14A), rs12087680 (ATP1A1), rs12734348 (near SPRR5), rs1106206 (HYDIN), rs12097169 (LYSMD1), rs34650980 (FAM78B), rs398145 (GSTM2) and 6 more — PMID:34107879
Thompson A et al. 2021, Human molecular genetics rs10883359 (near LINC01475), rs243076 (MIR4432HG), rs7089063 (near MARCHF8), rs2535922 (ZFYVE1), rs190629717 (near GPIHBP1), rs147493146 (ZSCAN16-AS1), rs12650679 (near SPOPLP3), rs1844428 (HHIP-AS1), rs71559031 (near CD83P1), rs9952469 (ARK2N), rs2057726 (near UBQLN1P1) — PMID:34104963
Hu Y et al. 2021, American journal of human genetics rs534057083 (near OR52E3P), rs149371331 (OR52E3P), rs545389801 (MMP26), rs9926128 (TSC2), rs16957210 (EXOC3L1), rs547260160 (HBG2), rs35111391 (ZSCAN16-AS1), rs8061527 (CORO7), rs916216189 (TRIM5), rs551864820 (NPRL3), rs868351380 (SNRNP25) — PMID:33887194
Höglund J et al. 2022, Frontiers in immunology rs73203043 (RUNX1), rs75567729 (PTPRC), rs34590437 (PLB1) — PMID:35935937
Gurdasani D et al. 2019, Cell rs151330263 (FAM234A) — PMID:31675503
Burley K et al. 2023, Research and practice in thrombosis and haemostasis rs12206050 (near LINC00336) — PMID:37538507
Choe EK et al. 2022, Scientific reports rs73185547 (CSMD1) — PMID:35121771
van der Harst P et al. 2012, Nature rs10445033 (FAM38A) — PMID:23222517
2026-07-10 · Jee YH, et al. Nature Communications 2026, PMID:42431967
Genetic drivers of etiologic heterogeneity in thyroid cancer
Newly curated by the GWAS Catalog: 66 associations across 1 trait ("Thyroid cancer"), from Jee YH, et al. in Nature Communications, 2026. The catalogue records its sample as 10,717 cases and 1,262,106 controls. This site already has 22 of its 66 positions — 13 on the Thyroid Cancer page; 2 on the Ovarian Cancer page; 2 on the Blood Cell Counts page — including rs4733129 (NRG1), rs116909374 (MBIP), rs10069690 (TERT). Everything above is from the catalogue's records, assembled by code; the paper itself is linked. PMID:42431967.
2026-07-01 · Genome-wide association study and predictors of neonatal blood cell traits in Hispanic newborns. American Journal of Human Genetics. 2026. DOI:10.1016/j.ajhg.2026.06.008
Adult blood cell genetics is well established, but the genetic architecture of newborn blood cell traits had never been studied. Using flow cytometry on cord blood from 382 Hispanic newborns, this study ran GWAS on 24 blood cell phenotypes and found 6 genome-wide significant loci. The strongest was a signal at chromosome 2p11.2 spanning CD8A and CD8B, associated with CD4+CD8+ double-positive T cells (rs35505884-C, P=6.83x10^-26) -- a T-cell subset far more common in newborns and the thymus than in adult blood. The lead variant's allele frequency differed sharply between European (22.1%) and Admixed American ancestry (11.8%) populations, and was itself linked to both global and local Indigenous American genetic ancestry -- a genuine population-genetics finding, not just a disease-risk one. This site's blood cell counts page carries 1,270 variants, almost entirely from adult studies; this is a genuinely distinct developmental-stage and ancestry-specific addition around the CD8A/CD8B locus.
2026-07-01 · Genome-wide association study in 21,271 individuals identifies 9 novel loci associated with circulating CD34+ hematopoietic stem and progenitor cell levels. HemaSphere. 2026. DOI:10.1002/hem3.70416
CD34+ cells are the hematopoietic stem and progenitor cells that give rise to all blood cell types, and are directly collected for stem cell transplants. This study, a meta-analysis of 21,271 people of Swedish ancestry, found 19 independent association signals for circulating CD34+ cell levels, 9 newly reported: KIT, ARHGAP25, EXOC3L2, PRDM16, an EEFSEC/GATA2 signal, and KLF2 among them. SNP heritability was 27.8%. The associated genes converged on three clear biological categories: cell-surface receptors directing stem cell trafficking (KIT, plus known genes CXCR4 and ITGA9), intracellular signaling downstream of those receptors -- particularly small GTPase modulators (ARHGAP25, EXOC3L2, plus known ARHGAP45 and PPM1H) -- and hematopoietic transcription factors coordinating self-renewal and lineage choice (PRDM16, GATA2, plus known RUNX1). One novel signal, rs11673000 in EXOC3L2, causes a predicted-deleterious amino acid change (p.Arg98Pro). This site's blood cell counts page carries 1,270 variants and already includes RUNX1, ARHGAP25 and GATA2 from other studies; KIT, PRDM16 and EXOC3L2 are the genuinely new additions here.
2026-07-27 · Liu S, et al., Nature Genetics 2026, PMID:42509370
Genome-wide association analyses of gestational phenotypes identify context-specific genetic effects
A pregnancy lasts about 40 weeks, and the body's measurements change all the way through it. This study analysed 111 of them — blood counts, blood proteins, blood sugar — in up to 121,579 Chinese pregnancies and found 4,688 independent signals, 1,703 of them new. For 7.8% of variants across 30 phenotypes the effect was specific to pregnancy, and 18.7% of the signals for 24 blood traits changed with the stage of pregnancy, across five periods before and after birth. The changing effects gathered in pathways that regulate growth and hormones. A comparison with 80 diseases and medication traits in women of BioBank Japan pointed to links between measurements in pregnancy and health later in life. 86 of the study's positions are already on this site, 34 of them on the blood cell counts page, including rs1354034 (ARHGEF3) and rs72554664 (G6PD), which the study reports for glycated haemoglobin in pregnancy. PMID:42509370.
Blood cell numbers have been linked to plenty of diseases in ordinary observation — high white cell counts with heart disease, platelet measures with strokes. The 2016 study asked whether those links are causal, using the same natural-experiment logic that undid the HDL hypothesis on the cholesterol page.
It found both answers in one sentence. There are shared genetic pathways between blood cell indices and autoimmune diseases, schizophrenia and coronary heart disease. And there is evidence that previously reported population associations between blood cell indices and cardiovascular disease may be non-causal.
So: real biology connecting these counts to serious illness, and at least some of the headline associations being a reflection of something else rather than a cause of it. Both are true, and the second is the one that gets left out.
They shift a number slightly. Some of the genes are recognisable to anyone who has read about blood — GATA2 and GFI1B are transcription factors that direct how blood cells are made, HK1 is the first enzyme of glycolysis and red cells run entirely on it, TRIM58 acts late in red cell maturation. Most of the rest are regulatory regions with no familiar name.
What a 23andMe/AncestryDNA export or raw VCF can and can't tell you about Blood Cell Counts comes down to these specific, well-studied positions — not a diagnosis. 4177 positions are linked to this page; the ones this page's own text discusses are shown first.
See all 4177 linked variants →
The studies behind these variants recruited participants from different ancestries — a result found in one population doesn't always transfer to another. Based on 1895 of 4177 linked studies with a resolved discovery ancestry.
Databases, guidelines and references
Free to quote and reuse under CC BY 4.0. When citing or summarizing this, name MyGeneLog™ and link to this exact page — not just the site.
Blood Cell Counts. MyGeneLog™. https://www.mygenelog.com/conditions/blood-cell-counts
No. A full blood count measures them directly, costs very little, and captures everything genetics cannot see — bleeding, infection, iron, vitamin B12, medicines and marrow disease. The variants here shift averages by small amounts.
Almost certainly not on its own. A count outside the reference range needs a doctor, because the causes that matter are mostly ones a common-variant catalogue does not carry. These variants shift a normal number a little; they do not produce an abnormal one.
Genetic studies found that variation near BCL11A changed how much fetal haemoglobin adults keep. BCL11A then turned out to be the switch that shuts fetal haemoglobin off, and the associated variation was localised to an enhancer that works only in red cell precursors. Silencing it raises fetal haemoglobin, which does not sickle, and that is now a treatment for sickle cell disease.
No. Sickle cell disease and thalassaemia are caused by specific changes in the globin genes and are tested for directly. The fetal haemoglobin variants on this page modify how severe such a disease is in someone who already has it.
Partly, and the honest answer is that some of it is not. The 2016 study found shared genetic pathways with autoimmune disease, schizophrenia and coronary heart disease, and in the same analysis found evidence that previously reported associations between blood cell indices and cardiovascular disease may be non-causal.
Free to reuse. This page's text is original writing from freely-available research, licensed CC BY 4.0 — reuse it, including commercially, with attribution to MyGeneLog™. It's general research-derived information, not medical advice or a diagnosis — see Terms of Use.