Chromosomes, linkage and gene mapping

In short

A chromosome is one long DNA–protein thread carrying hundreds to thousands of genes in a fixed linear order, so genes on the same chromosome tend to be inherited together — this is linkage, the first major exception to independent assortment. Crossing over at meiosis breaks that togetherness in a fraction of gametes, and because that fraction grows with the physical distance between two loci, it can be used as a ruler: 1% recombinants = 1 centimorgan. Gene mapping — two-point and three-point test crosses, interference, mapping functions, LOD scores and marker-based maps — is the core quantitative content of classical genetics and is examined heavily in CSIR-UGC NET Life Sciences, GATE BT/XL, ICAR ASRB NET, DBT-BET and ICMR JRF. In practice it underpins marker-assisted breeding at ICAR institutes, linkage and association studies at CSIR-IGIB and NIPGR, and seed-industry trait pyramiding.

Explained

Chromosomes as the physical carriers of genes

• Chromosome: a single DNA molecule with histone and non-histone proteins; visible as a condensed body at metaphase. Haploid number n = number of linkage groups.
• Landmarks: centromere (kinetochore attachment, divides the chromosome into short arm *p* and long arm *q*), telomeres (repeat-capped ends), nucleolar organizer region / satellite (SAT) chromosome, constrictions.
• Morphological types by centromere position (Levan's arm-ratio system): metacentric (≈1.0–1.7), submetacentric (1.7–3.0), acrocentric (3.0–7.0), telocentric (>7.0).
• Chromosome theory of inheritance (Sutton and Boveri, 1902–03): genes are on chromosomes, because chromosome behaviour at meiosis parallels Mendelian segregation. Proved by Morgan's white-eye work in *Drosophila* (1910) and Bridges' non-disjunction evidence (1916).
• Haploid numbers to remember: human n = 23 (22 autosomes + X; Y carries few mapped genes), *Drosophila melanogaster* n = 4, garden pea n = 7, maize n = 10, rice n = 12, bread wheat n = 21 (hexaploid, 3 genomes), *Arabidopsis* n = 5, *Neurospora* n = 7.
• Special chromosomes used in mapping: polytene chromosomes of *Drosophila* salivary glands give reproducible bands, allowing a cytological map to be compared with the genetic map; lampbrush chromosomes show transcription in oocytes.

Linkage: definition, discovery and types

• Linkage: the tendency of genes located on the same chromosome to be transmitted together, giving an excess of parental (non-recombinant) gametes over the 1:1:1:1 expected from independent assortment.
• Bateson and Punnett (1906), sweet pea, flower colour and pollen shape: F2 deviated from 9:3:3:1 with excess of parental combinations. They called it "gametic coupling" and "repulsion".
• Morgan (1911) explained both as one phenomenon: coupling (cis) = dominant alleles entered from the same parent, written AB/ab; repulsion (trans) = AB entered separately, written Ab/aB. The linkage phase must be known before progeny are classified.
• Morgan's law of linkage: genes on the same chromosome are inherited together; the strength of linkage depends on the distance between them.
• Complete linkage: no crossover between the loci, only two gamete classes, both parental (for example loci very close together, or all genes in *Drosophila* males, where meiotic crossing over does not occur).
• Incomplete (partial) linkage: the usual case — all four classes appear, parental types in excess, recombinants < 50%.
• Pseudolinkage is the apparent linkage of unlinked genes caused by a translocation heterozygote; linkage disequilibrium is the population-level, non-random association of alleles and is not the same as linkage.
• Significance of linkage: keeps favourable allele blocks together (useful in breeding), restricts the variability available for selection, and is the basis of all mapping.

Crossing over: mechanism and evidence

• Crossing over: reciprocal exchange of corresponding segments between non-sister chromatids of homologous chromosomes, occurring at the pachytene stage of meiosis I within the synaptonemal complex; seen cytologically as a chiasma at diplotene.
• It happens at the four-strand (tetrad) stage, so a single crossover converts a tetrad into 2 parental + 2 recombinant chromatids — hence a single exchange in 100% of cells yields only 50% recombinants, and RF can never exceed 50%.
• Molecular basis: SPO11-mediated double-strand break → strand invasion → Holliday junction → resolution; the double-strand-break repair (Szostak) model explains both crossovers and gene conversion. Exchange is not random along the chromosome — there are recombination hotspots (PRDM9-directed in humans) and cold regions near centromeres and heterochromatin.
• Types: single, double, multiple crossovers; reciprocal (equal) versus unequal; somatic/mitotic crossing over (rare, gives twin spots); sister chromatid exchange produces no new allele combinations.
• Evidence that genetic recombination equals physical exchange: Stern (1931) in *Drosophila* using attached-X and marker chromosomes; Creighton and McClintock (1931) in maize, where recombinant progeny also carried the exchanged cytological knob and translocated segment.
• Factors affecting crossover frequency: distance between loci; proximity to centromere/telomere (reduced); sex (higher in human females, nil in *Drosophila* males); age; temperature and radiation (increase it); chromosomal rearrangements — an inversion heterozygote suppresses recovery of recombinants (balancer chromosomes exploit this).

Recombination frequency and two-point mapping

• Recombination frequency (RF) = recombinant progeny / total progeny × 100. 1% recombinants = 1 map unit = 1 centimorgan (cM); 100 cM = 1 Morgan.
• Best measured in a test cross (F1 heterozygote × homozygous recessive), because every gamete's genotype is read directly in the phenotype. In F2 self/intercross data RF must be estimated by maximum likelihood or product-ratio methods — do not count phenotypes directly.
• Steps for a two-point test cross: write the F1 genotype with its phase → list the four progeny classes → identify the two parental classes (the most frequent, matching the F1 input combinations) → the other two are recombinants → RF = recombinants/total × 100 → test linkage by chi-square against 1:1:1:1 with 3 df.
• Interpretation: RF ≈ 50% and non-significant χ² → genes independent (different chromosomes, or far apart on the same one); RF significantly < 50% → linked.
• Sturtevant (1913) built the first genetic map, six X-linked *Drosophila* genes, using the additivity of map distances — if a–b = 5 and b–c = 8, then a–c ≈ 13.
• Limitation: RF underestimates distance beyond about 20–25 cM because double crossovers restore the parental arrangement and go uncounted; the curve of RF against true distance flattens at 50%.

Three-point test cross, gene order and interference

Three loci in one cross give order, two distances and interference at once. Eight phenotypic classes appear in four reciprocal pairs.

1. Identify the parental (non-crossover) pair — the two most frequent classes; they reveal the F1 phase.
2. Identify the double-crossover (DCO) pair — the two rarest classes.
3. Determine the middle gene: compare a parental class with a DCO class; the single gene whose allele has "switched" is the middle gene (a DCO flips only the central marker).
4. Rewrite all classes in the correct order, then assign each pair to region I (left–middle), region II (middle–right) or DCO.
5. RF (region I) = (SCO I + DCO)/total × 100; RF (region II) = (SCO II + DCO)/total × 100. Always add the DCOs — they are crossovers in both regions.
6. Total map length = region I + region II; the directly observed outer-marker RF is smaller by 2 × DCO%.
7. Interference: expected DCO = RF(I) × RF(II) × total. Coefficient of coincidence (CoC) = observed DCO/expected DCO; interference I = 1 − CoC.
8. Positive interference (CoC < 1) is the norm — one crossover reduces the chance of a second nearby; it falls towards zero as the regions get farther apart. Negative interference (CoC > 1) is rare and usually indicates clustering or small samples. Complete interference gives CoC = 0.
9. The coincidence/interference value is not a probability of linkage; it is a measure of crossover independence.

Mapping functions, and other routes to a map

• Mapping functions convert an observed RF (r) into additive map distance (d) by modelling multiple crossovers: Haldane (1919) assumes no interference (Poisson); Kosambi (1944), by the Indian mathematician D. D. Kosambi, allows interference and is the function normally used in plant and animal linkage software (MAPMAKER, JoinMap, QTL IciMapping).
• Tetrad analysis (*Neurospora*, *Saccharomyces*): ordered octads give centromere-to-gene distance from second-division segregation — map distance = ½ (number of MII tetrads)/total × 100; unordered tetrads use parental ditype (PD), non-parental ditype (NPD) and tetratype (T): RF = (T/2 + 3NPD)/total × 100; PD ≈ NPD indicates no linkage.
• Bacterial and phage mapping uses interrupted mating/time-of-entry (conjugation), co-transduction and co-transformation frequencies, not cM.
• Human mapping: pedigree linkage analysis with LOD scores (Morton, 1955) — LOD ≥ +3 accepted as linkage, ≤ −2 as exclusion; parametric linkage for Mendelian disorders, non-parametric sib-pair sharing and genome-wide association (GWAS) for complex traits.
• Somatic cell hybridisation (human–rodent hybrids) and *in situ* hybridisation/FISH assign genes to chromosomes and bands — these give physical maps, measured in base pairs.
• Molecular marker linkage maps: RFLP → RAPD/AFLP → SSR (microsatellite) → SNP arrays and genotyping-by-sequencing; mapping populations are F2, backcross, recombinant inbred lines (RILs), doubled haploids and near-isogenic lines. QTL mapping (interval mapping, composite interval mapping) locates quantitative trait loci between flanking markers.
• Genetic versus physical distance: the human genome is about 3,100 Mb of DNA with a sex-averaged genetic map of roughly 3,500–3,800 cM, so 1 cM ≈ 0.8–1 Mb on average — but the ratio varies several-fold between hotspots and centromeric cold spots, so cM cannot be converted to bp locally.
• Indian practice and regulation: ICAR institutes use marker-assisted selection and backcrossing routinely (bacterial-blight genes *Xa21*, *xa13*, *xa5* pyramided into Improved Samba Mahsuri and Pusa Basmati lines; the *Sub1* QTL in Swarna-Sub1 for submergence tolerance). DBT funds the GenomeIndia whole-genome resource; human genetic studies require Institutional Ethics Committee clearance under the ICMR National Ethical Guidelines for Biomedical and Health Research (2017) and the New Drugs and Clinical Trials Rules, 2019; prenatal use of genetic tests is controlled by the PCPNDT Act, 1994. A dedicated DNA technology regulation Bill was introduced but withdrawn, so there is no separate DNA profiling statute at present.

Key points

• Linkage – genes on the same chromosome inherited together; parental gamete classes in excess.
• Linkage group – all genes on one chromosome; number of groups = haploid chromosome number.
• Coupling (cis) / repulsion (trans) – the two possible phases of a double heterozygote, AB/ab and Ab/aB.
• Complete linkage – only parental gametes, RF = 0; incomplete linkage – RF between 0 and 50%.
• Crossing over – reciprocal exchange between non-sister chromatids at pachytene; chiasma is its cytological image.
• Four-strand stage – exchange involves 2 of 4 chromatids, so maximum RF is 50%.
• Recombination frequency – recombinants/total × 100; measured in a test cross.
• Centimorgan (map unit) – 1% recombination; 100 cM = 1 Morgan.
• Additivity – map distances add along a chromosome; observed RF does not, beyond ~25 cM.
• Double crossover – rarest class in a three-point cross; restores parental outer markers and switches the middle gene.
• Gene order rule – compare parental with DCO class; the gene that differs is in the middle.
• Interference – I = 1 − CoC; usually positive, so observed DCOs are fewer than expected.
• Mapping functions – Haldane (no interference) and Kosambi (with interference) convert RF to additive distance.
• Tetrad analysis – RF = (T/2 + 3NPD)/total × 100; PD ≈ NPD means no linkage.
• LOD score – log10 likelihood ratio; ≥ 3 proves linkage, ≤ −2 excludes it.
• Genetic map ≠ physical map – cM reflects crossover frequency, which varies with hotspots, sex and centromere proximity.
• Creighton and McClintock (1931) – maize proof that genetic recombination accompanies physical chromosome exchange.
• Inversion heterozygotes and *Drosophila* males – crossover products not recovered, so no mapping is possible there.

Formulas and examples

Example 1 — two-point test cross and a test for linkage. *Drosophila* F1 females AB/ab crossed to ab/ab males; 1,000 progeny: AB 420, ab 400, Ab 90, aB 90.

• Parental classes AB and ab (phase was cis) → recombinants = 90 + 90 = 180.
• RF = 180/1000 × 100 = 18%, so a–b = 18 cM.
• χ² against 1:1:1:1 (E = 250 each) = (170)²/250 + (150)²/250 + 2 × (160)²/250 = 115.6 + 90.0 + 204.8 = 410.4, df = 3, p ≪ 0.001 → the genes are linked.

Example 2 — three-point test cross: order and distances. F1 heterozygote test-crossed; 1,000 progeny scored for a, b, c:

• Parentals = *a b c* / *+ + +*; DCOs = *a + c* / *+ b +*. Comparing *a b c* with *a + c*, only b has switched → b is the middle gene; order is a – b – c.
• RF(a–b) = (43 + 43 + 7 + 7)/1000 × 100 = 100/1000 = 10 cM.
• RF(b–c) = (93 + 93 + 7 + 7)/1000 × 100 = 200/1000 = 20 cM.
• Map: a —10 cM— b —20 cM— c, total 30 cM.

Example 3 — interference, and why the outer distance looks short (same data).

• Expected DCO = 0.10 × 0.20 × 1000 = 20; observed = 14 → CoC = 14/20 = 0.70, interference I = 1 − 0.70 = 0.30 (30% positive interference).
• Progeny recombinant for the outer pair a–c = 86 + 186 = 272 → observed RF(a–c) = 27.2%, not 30%, because the 14 DCOs are parental for a and c.
• Check the correction: 27.2 + 2 × 1.4 = 30.0 cM ✓ — this is why three-point data are preferred to three separate two-point crosses.

Example 4 — mapping functions at high RF. Two markers show r = 0.45 (45% recombinants).

• Haldane: d = −0.5 × ln(1 − 0.90) = −0.5 × ln(0.10) = −0.5 × (−2.3026) = 1.1513 M = 115.1 cM.
• Kosambi: d = 0.25 × ln(1.90/0.10) = 0.25 × ln(19) = 0.25 × 2.9444 = 0.7361 M = 73.6 cM.
• The raw 45 cM badly underestimates distance; Kosambi gives the smaller correction because it allows interference. Mapping-function choice must be stated with any published map.

Example 5 — LOD score in a human pedigree. Pooling five phase-known families: 40 informative meioses, 8 recombinants, so the maximum-likelihood θ̂ = 8/40 = 0.20.

• L(θ = 0.20) = 0.20⁸ × 0.80³² = 2.56 × 10⁻⁶ × 7.9228 × 10⁻⁴ = 2.0282 × 10⁻⁹.
• L(θ = 0.50) = 0.5⁴⁰ = 9.0949 × 10⁻¹³.
• Ratio = 2230.1 → LOD = log10(2230.1) = 3.35 > 3, so linkage at about 20 cM is accepted. Had only 10 meioses with 2 recombinants been scored, LOD would be 0.84 — the same θ̂, but no evidence.

Where it is used

• Plant breeding (ICAR, state agricultural universities, private seed companies): marker-assisted backcrossing and trait pyramiding in rice, wheat, maize, cotton and pulses — bacterial-blight resistance genes in Improved Samba Mahsuri, the *Sub1* submergence QTL in Swarna-Sub1 — where tight marker–gene linkage replaces slow phenotyping.
• Animal breeding and dairy genetics (NDRI, NBAGR): linkage and SNP-based genomic selection for milk yield and disease resistance in cattle and buffalo.
• Clinical and medical genetics: linkage analysis and haplotype-based carrier/prenatal testing for monogenic disorders (thalassaemia, Duchenne muscular dystrophy, spinocerebellar ataxias) at CMC Vellore, AIIMS, SGPGI and CSIR-IGIB; informed consent and ethics clearance under ICMR guidelines, prenatal use governed by the PCPNDT Act, 1994.
• Bioinformatics and genomics roles: building and curating linkage/QTL maps, anchoring genome assemblies to linkage groups, recombination-rate estimation, LOD and GWAS pipelines (R/qtl, PLINK, MERLIN, TASSEL) — hiring at Strand Life Sciences, MedGenome, Eurofins, Novartis/Syngene, Bayer and Corteva research centres in Bengaluru and Hyderabad.
• Forensic and identity testing (CFSLs, state FSLs): STR loci are deliberately chosen on different chromosomes or far apart so they segregate independently; linked markers would violate the product rule used to compute match probabilities.
• Teaching and research laboratories: *Drosophila* and maize three-point crosses remain the standard B.Sc/M.Sc practical, and *Neurospora* tetrad analysis the standard centromere-mapping exercise.

Interview questions

• Q: Why can recombination frequency never exceed 50%? A: A crossover occurs at the four-chromatid stage and converts only two of the four chromatids, so even a crossover in every meiosis yields 50% recombinants; higher values mean the loci are effectively unlinked.
• Q: How do you decide the middle gene in a three-point test cross? A: Compare a parental class with a double-crossover class — the one gene whose allele differs between them is the middle gene, because a double crossover only flips the central marker.
• Q: Why must double crossovers be added to both regional recombination frequencies? A: A double crossover represents an exchange in region I *and* in region II, so omitting it underestimates both distances and the total map length.
• Q: What is interference and what does a coefficient of coincidence of 0.7 mean? A: Interference is the effect of one crossover reducing the chance of a second nearby; CoC 0.7 means only 70% of the expected double crossovers occurred, so interference is 0.30.
• Q: Distinguish linkage from linkage disequilibrium. A: Linkage is the physical co-location of loci on one chromosome, measured in families as recombination fraction; linkage disequilibrium is the non-random association of specific alleles in a population, which decays over generations with recombination.
• Q: Why is a test cross preferred over an F2 for measuring RF? A: In a test cross each gamete from the heterozygote is expressed directly in the progeny phenotype, so recombinants can be counted; in an F2 the classes overlap and RF must be estimated statistically.
• Q: Which experiment proved that genetic recombination corresponds to physical chromosome exchange? A: Creighton and McClintock's 1931 maize work, in which recombinant progeny also carried the exchanged cytological markers (knob and translocated segment); Stern showed the same in *Drosophila*.
• Q: Why do genetic map distances not match physical distances in base pairs? A: Recombination rate varies along the chromosome — hotspots inflate cM while centromeric and heterochromatic regions are crossover-poor — so a cM corresponds to roughly 1 Mb only on average in humans.
• Q: Why can you not map genes using *Drosophila* males? A: Meiotic crossing over does not occur in *Drosophila* males, so no recombinants are produced; mapping crosses always use heterozygous females.
• Q: How does an inversion heterozygote affect mapping? A: Crossover products within the inversion loop are duplicated/deficient and inviable, so recombinants are not recovered and the region appears falsely tightly linked — the basis of balancer chromosomes.
• Q: When would you use the Kosambi rather than the Haldane mapping function? A: When interference is present, which is typical in eukaryotes; Haldane assumes a Poisson, interference-free process and therefore inflates distances at high recombination fractions.
• Q: What LOD score is accepted as evidence of linkage, and why that threshold? A: LOD ≥ +3, i.e. odds of 1,000:1 in favour of linkage, which approximately corresponds to a genome-wide significance of about 5% for a single Mendelian trait; LOD ≤ −2 excludes linkage.

Common mistakes

• Forgetting to add the double-crossover classes when calculating each regional recombination frequency, which shortens the map.
• Assuming the parental classes are the two with the dominant or "wild-type" phenotypes instead of reading them from the actual F1 phase — repulsion-phase crosses are then scored backwards.
• Using F2 phenotype counts directly as recombinant frequencies, or counting both reciprocal recombinant classes but dividing by the wrong total.
• Treating observed RF as additive over long distances, so that a–c comes out less than (a–b) + (b–c) and the discrepancy is called an error rather than double crossing over.
• Confusing interference with the coefficient of coincidence (they are complements, I = 1 − CoC), or reporting a negative interference value from a handful of progeny.
• Saying crossing over occurs between sister chromatids, or at the two-strand stage — it is between non-sister chromatids of homologues at pachytene.
• Equating cM with kilobases, or quoting a genetic map without naming the mapping population and mapping function used.

Related topics

Genetics · Mendelian inheritance: gene interaction and pedigree analysis · Cytogenetics and chromosomal disorders · Cell biology · Molecular markers and QTL mapping · Population genetics and Hardy-Weinberg equilibrium