Guide to the Human Genome
Home | Table of Contents | Search text | Search genes | Search sequences | Purchase | FAQ | Blog | Help

Gene Structure

The tables and figures in this section use a different gene set than that used in most of the text. The set used here was chosen because it derives from the set of transcripts mapped onto the human genome reference DNA sequence. Some of the differences between the mapped transcripts and the cDNA sequences used to predict the RefSeq proteins derive from some members of the latter set not mapping precisely onto the reference genome sequence. A single transcript from each named gene was used to produce the tables and figures in this section. Transcript predictions and some other special cases were excluded (see Notes and References). The retained set was 18,159 transcripts.

Gene sizes

Human genes vary over a wide size range. This is illustrated in the following figure. The size shown is for a single transcript from each gene including its introns; alternate products, in some cases, may span a larger combined region. The plot is cumulative with the y-axis showing the percentile. About 15% of the transcripts span greater than 100 kb of genomic sequence. The median size was 23,329 nucleotides. Genes with no annotated UTR (excluded from this set) tend to be small so these values may be overestimates. These estimates are further complicated by issues relating to the size of the annotated UTRs.

Gene Sizes

The following table presents some of the largest genes in the human genome. These sizes are from transcribed regions rounded to the nearest 0.01 Mb. For comparison, genes encoding some of the largest proteins have been included. For more information about these genes, see the section listed in the right column. Some of these genes produce a very large number of transcripts and isoforms. Many have functions in the development of the nervous system. Note that links for the gene names point to a single isoform / transcript and that the reference set may include others.

Largest Genes in the Genome
GeneSize (Mb)ProteinSection
CNTNAP2 2.30 Caspr2 protein Neurons
PTPRD 2.30 receptor protein tyrosine phosphatase D Protein Tyrosine Phosphatases
DMD 2.22 dystrophin Muscle
DLG2 2.17 chapsyn-110 Synapses
CSMD1 2.06 Additional Interaction Domain Families
MACROD2 2.06 Additional Genes in Development
EYS 1.99 Crystallins and Other Eye proteins
LRP1B 1.90 lipoprotein receptor family Lipoproteins
CTNNA3 1.78 α catenin 3 Cadherins and Related Proteins
A2BP1 1.69 ataxin 2 binding protein Cerebellum
FHIT 1.50 dinucleoside triphosphate hydrolase Nucleotide Pathways
AGBL4 1.49 Carboxypeptidases
FAM190A 1.47
GPC5 1.47 glypican 5 Protein Glycosylation
GRID2 1.47 glutamate receptor Glutamate
NRXN3 1.46 neurexin 3 Neurons
MAGI2 1.44 membrane guanylate kinase PDZ Domain
DPP10 1.40 dipeptidyl peptidase family Serine Proteases
PARK2 1.38 parkin Neurons
IL1RAPL1 1.37 receptor accessory protein Interleukins and Their Receptors
CNTN5 1.34 contactin 5 Neurons
PRKG1 1.30 protein kinase Cyclic Nucleotides
DAB1 1.25 D. melanogaster disabled homolog 1 Additional Membrane Functions
ANKS1B 1.25 cajalin 2 Nucleus and Nucleolus
GALNTL6 1.23 Protein Glycosylation
KCNIP4 1.22 Potassium Channels
CSMD3 1.21 Additional Interaction Domain Families
IL1RAPL2 1.20 receptor accessory protein Interleukins and Their Receptors
AUTS2 1.19 Fibroblast Growth Factors
DCC 1.19 netrin receptor Netrins and Laminins
GPC6 1.18 glypican 6 Protein Glycosylation
CDH13 1.17 cadherin 13 Cadherins and Related Proteins
ERBB4 1.16 EGF receptor family Epidermal Growth Factor
SGCZ 1.15 sarcoglycan zeta Muscle
ACCN1 1.14 cation channel Sodium Channels
CTNNA2 1.14 α catenin 2 Cadherins and Related Proteins
SPAG16 1.13 sperm antigen Testes and Sperm
NRG1 1.12 neuregulin 1 Neurons
OPCML 1.12 Neurons
PTPRT 1.12 protein tyrosine phosphatase Protein Tyrosine Phosphatases
NRXN1 1.11 neurexin 1 Neurons
WWOX 1.11 oxidoreductase WW Domain
NRG3 1.11 neuregulin 3 Neurons
CDH12 1.10 cadherin 12 Cadherins and Related Proteins
DPP6 ~ 1.10 dipeptidyl peptidase family Serine Proteases
PARD3B 1.07 tight junction protein PDZ Domain
PTPRN2 1.05 protein tyrosine phosphatase Protein Tyrosine Phosphatases
SOX5 1.03 transcription factor SOX Family
NKAIN2 1.02 T cells
Genes for Large Proteins
GeneSize (Mb)ProteinSection
TTN 0.28 titin Muscle
MUC16 0.13 mucin 16 (CA-125 antigen) Mucins

TTN and MUC16 are the largest proteins in the reference set but their genes are only a fraction of the size of the largest genes.

As can be seen in the following figure, in general, these large genes are dispersed along the chromosomes; however, SPAG16 and ERBB4 are very close to each other on chromosome 2. GPC5 and GPC6 are near each other on chromosome 13. Note the absence of large genes on the gene-rich chromosomes 19 and 22.

Chromosomal Locations of Large Genes

Related proteins are sometimes encoded by genes that have very different sizes. Although utrophin (UTRN) is encoded by a large gene (0.56 Mb), it is only a fraction of the size of dystrophin (DMD, 2.22 Mb). DAB2 is a 0.05-Mb gene, much smaller than DAB1 (1.25 Mb). LRP1 (0.08 Mb) is also much smaller than LRP1B (1.9 Mb).

As seen in the preceding table, two of the neurexins are encoded by very large genes but the third family member, NRXN2, is only 0.12 Mb. A similar situation is found with the roundabout (ROBO) family and several other neuronal protein families (see Neurons). The SNRPN gene in the Prader–Willi imprinted region and the SNRPB gene (see Capping and Splicing) also differ greatly in size but encode similar-sized proteins.

Alternate transcripts and isoforms

A number of genes in the reference set produce a large number of distinct transcripts, often leading to a similarly large number of isoforms. The neurexins (see Neurons) are encoded by extremely large genes that produce an exceptional number of isoforms via alternate splicing. Other genes with very large numbers of transcripts include CMTM1 (chemokine-like protein), COL13A1 (type XIII collagen), CREM (cAMP response modulator), DMD (dystrophin), and PDE9A (cGMP phosphodiesterase).

Alternate transcripts are a mechanism for producing isoforms targeted to distinct subcellular compartments. Isoforms are also produced in a tissue-specific manner. HK1 (see Hexokinases and Initial Sugar Metabolism) produces several isoforms from different transcripts, some of which are testes-specific.

Exon / intron structure

Human genes vary widely in the number and size of their exons and introns. The different sequences found at exon / intron junctions are detailed in the section on Capping and Splicing.

The following figure shows the distribution of exon number for human genes. The number of genes with a given exon count is the y coordinate. It uses the gene set described at the beginning of this section. The distribution has a mode of four exons and a median of eight exons. The small number of genes with over 100 exons (see table later in this section) is not plotted.

Exon Numbers in Human Genes

Many genes are interrupted by an extremely large number of introns. The following table presents some of them. Note that these are not the largest genes in the genome, but they encode many of the largest proteins. Only one transcript from each gene was used. The total number of exons for the gene may be larger than shown. Also, not all transcripts from these genes may be present in the current data sets. The links in the table point to the isoform / transcript with the indicated number of exons.

Genes with the Most Exons
GeneExon countProteinSection
TTN 312 titin Muscle
NEB 150 nebulin Muscle
SYNE1 146 nesprin 1 Spectrin and Plectin Families
COL7A1 118 collagen type VII Collagen
SYNE2 116 nesprin 2 Spectrin and Plectin Families
HMCN1 107 hemicentin 1 Additional Immunoglobulin-related Receptors
RYR1 106 skeletal muscle ryanodine receptor Muscle
UBR4 106 retinoblastoma-associated protein RB1 and Related Functions
OBSCN 106 obscurin Muscle
RYR2 105 cardiac muscle ryanodine receptor Muscle
RYR3 104 ryanodine receptor Muscle
SSPO 103 subcommissural organ spondin Additional Genes in Development
MDN1 102 midasin Nucleus and Nucleolus

Many proteins are encoded by genes with a single exon or have multiple exons but no introns in their protein-coding regions. Examples are found in the histones, the olfactory and other G-coupled receptors, the interferons, and some members of the FOX family. As seen in the preceding table, large proteins are generally encoded in genes interrupted by many introns. A notable exception is EPPK1 (epiplakin, a protein of over 5000 amino acids) which may lack introns in most or all of its coding sequence.

The following figures show how exon number correlates more with protein size than gene size, notably for genes with many exons.

Exon Count vs Protein Size Exon Count vs Gene Size

The plot on the left has protein size (log scale) on the x-axis. Gene size (log scale) is the x-axis in the plot at right. Exon number is given on the y-axis (log scale). Single-exon genes are the points along the x-axis. Note the differing scales on the x-axes. The log scales help present the wide data range. The gene set used here is the same as that used in the figure on exon numbers for human genes. Gene size is the span of the transcribed region. The UTRs may be underestimated (see below).

The final plot in this series presents gene size against protein size. A positive correlation is observed.

Gene Size vs Protein Size

The same gene / transcript set was used as in the previous figures. The roughly linear set of points at the bottom of the cluster derives from single-exon genes with very small reported UTRs.

Introns vary over a very large size range. The following table uses the same gene set used to produce the figures on exon numbers to present median intron sizes. The table shows data for genes with 2 through 16 exons (1 through 15 introns). The "Gene count" column is the number of examples of that type. Note the greatly increased size for the first introns of genes compared to their subsequent introns and the inreasing size of first and other early introns for genes with many exons.

Median Intron Sizes

The following table lists some of the largest documented introns in the genome. Very large introns are, by necessity, found in large genes. This list overlaps with the list of the largest genes earlier in this section. Note how the genes with the largest introns vary considerably in the number of introns they contain. DPP6, a very large gene spanning an assembly gap, also is likely to contain a very large intron. Many of the genes listed in this table have multiple entries in the reference set for distinct isoforms and transcripts. The links in the following table point to the isoform / transcript with the indicated large intron.

Genes with the Largest Introns
size (bp)
Intron countLargest intron (bp)ProteinSection
KCNIP4 1,220,136 7 1,097,903 Kv channel interacting protein Potassium Channels
ACCN1 1,143,721 9 1,043,911 cation channel Sodium Channels
NRG1 1,103,504 4 955,100 neuregulin 1 Neurons
DPP10 1,402,038 25 866399 dipeptidyl peptidase family Serine Proeases
WWOX 1,113,014 8 778,855 oxidoreductase WW Domain
LRRTM4 774,654 3 769,401 Neurons
HS6ST3 748,720 1 740,920 heparan sulfate sulfotransferase Protein Glycosylation
GPC5 1,468,556 7 721,292 glypican 5 Protein Glycosylation
SGCZ 1,148,420 7 682,658 sarcoglycan zeta Muscle
PDE4D 924,757 14 677,200 cAMP phosphodiesterase Cyclic Nucleotides
CNTNAP2 2,304,634 23 657,297 Caspr2 protein Neurons
FAM155A 698,205 2 654,926
PCDH9 927,503 3 593,993 protocadherin 9 Cadherins and Related Proteins
OPCML 1,117,529 7 589,253 Neurons
DLG2 2,172,260 27 576,930 chapsyn 110 Synapses
RORA 741,020 10 550,366 RAR-related receptor Nuclear Receptors
MACROD2 2,057,697 16 544,980 Additional Genes in Development
NTM 966,346 7 540,674 neurotrimin Neurons
IL1RAPL2 1,200,827 10 536,480 receptor accessory protein Interleukins and Their Receptors
FGF14 680,920 4 526,174 fibroblast growth factor 14 Fibroblast Growth Factors
IMMP2L 899,238 5 523672 Mitochondria
FHIT 1,502,098 9 522,714 dinucleoside triphosphate hydrolase Nucleotide Pathways
FAM190A 1,474,687 10 512,577
ODZ2 979,320 28 500,512 Additional Brain Proteins

The following table gives exon size data for genes with up to 15 exons using the same genes set desribed for the corresponding intron table. The sizes of the first exons are likely underestimated because of incomplete cDNA clones. The sizes of the final exons are likely overestimated because longer mRNAs are often mapped onto the genome. They may include other poly(A) processing sites that would result in shorter mRNAs. Middle exons have a relatively consistent median size. This number declines modestly as number of exons in the transcript increases. For all middle exons from the full set of selected transcripts, the median value is 123 nucleotides.

Median Exon Sizes
Exons Gene

Gene size and gene density

As described in the section on the chromosomes, there is considerable variation in gene density among the chromosomes. The following figure uses the same gene set used for the exon correlation figures earlier in this section to examine how gene size varies on the chromosomes (except that the genes in the pseudoautosomal regions of the X and Y were included in the data for both chromosomes instead of being used just once). Median gene size in kb is plotted against kb of sequenced DNA per gene (the x-axis being a reciprocal measure of gene density).

Gene Size and Gene Density

A clear trend is seen where increased gene density (shown here as less sequenced DNA per gene) is associated with a decrease in gene size. The increased gene density is not simply similar-sized genes being closer to each other. The genes still occupy only a fraction of the DNA of the chromosomes (even if predicted genes were added to the set). The X and Y chromosomes and the autosomes with the highest and lowest gene densities are labeled. A notable exception to the trend is seen with the Y chromosome (and to a lesser degree with the X chromosome). The Y chromosome has relatively few genes compared to the other chromosomes.

CpG islands

Although the CpG dinucleotide is generally found at lower frequencies than expected based on base composition, there are regions of the genome called CpG islands where the frequency of this dinucleotide is relatively higher. CpG islands are often located near the starts of mRNAs.

CpG Islands and mRNA Starts

The preceding figure is a cummulative plot of the distances from the center of the nearest CpG island to 5' ends of the transcripts mapped onto the reference genome (selected as before but also excluding genes assigned to chromosome fragments with no CpG island). Negative distances indicate an upstream relative location. Almost 61% of the selected transcripts have their starts within a CpG island. As can be seen in the figure, an even higher fraction of genes has RNA starts close to strictly defined CpG islands. The distribution has very long tails. For comparison, the equivalent calculation for mRNA 3' ends has a relatively flat distribution (not shown).

Notes and references

Many references and other information for individual genes can be found in the RefSeq entries linked via the pages for the proteins mentioned in this section. A table of these entries (with the corresponding gene identifiers) and a collection of their sequences also are available.

The gene information for this section is based on the release 37.1 reference genome sequence and the NCBI Map Viewer tables. The size of DPP6 is an estimate as it spans a gap in the assembly. The CpG islands used to prepare the figure were those defined as "strict" in the genome annotation.

The transcript set used to prepare the figures and tables was constructed from the set of transcripts in the Map Viewer tables. For each named gene, only one transcript with a largest encoded protein (in amino acids) was used. If a gene had multiple transcripts encoding proteins of that size, one with the most exons was retained. Transcript predictions were excluded. Similarly, genes reported with no untranslated region were also generally excluded (many of these were olfactory receptor genes). The retained set had 18,159 transcripts. It also excluded a small number of ambiguously placed transcripts and a few genes that span gaps in the assembly.

See also the additional reading for this chapter.

Previous section | Additional reading | Next section

Home | Table of Contents | Search text | Search genes | Search sequences | Purchase | FAQ | Blog | Help

Guide to the Human Genome
Copyright © 2010 by Stewart Scherer. All rights reserved.

CSHL Press