Written and maintained by CASRAI Editorial Board
Last updated
Short answer. Recombinant inbred lines (RILs) are sets of fully inbred, homozygous lines that all descend from a cross between two inbred parents, so each line carries its own fixed mosaic of parental chromosome segments. Because the lines are fixed, researchers can phenotype the same genotypes repeatedly and use them for QTL mapping. In plants they are made by repeated selfing; in mice and rats by 20 or more generations of brother x sister mating.
What recombinant inbred lines are
The Mouse Genome Informatics (MGI) nomenclature rules describe recombinant inbred (RI) strains as strains that contain “unique, approximately equal proportions of genetic contributions from two original progenitor inbred strains.” Each line in a set is therefore a different recombinant of the same two parents, and, once inbred, it breeds true. A panel of such lines is a reusable mapping resource: because the genotypes are fixed, the same population can be replicated and analysed repeatedly (Keurentjes et al.).
The term is used for plants (Arabidopsis, maize, rice and other crops, where it often appears as “recombinant inbred lines plant breeding”) and for the mouse and rat RI “strains” in the MGI vocabulary. The genetics is the same; only the breeding scheme and naming conventions differ.
How RILs are produced
In plants: repeated selfing
Keurentjes et al. (open access in PMC) describe RILs as homozygous lines obtained by repeated selfing of segregating progeny. Their Arabidopsis comparison used 161 Cvi x Ler RILs advanced to the F13 generation. Residual heterozygosity in that set was estimated at 0.71% in the F10 generation; the lines were genotyped again at F13 and confirmed as practically 100% homozygous.
In mice and rats: sibling mating
MGI states that RI strains are traditionally formed by crossing two inbred strains and then carrying out 20 or more consecutive generations of brother x sister matings. The same page notes an alternative: starting from advanced intercross lines, where F2 animals are mated without sib-mating for several generations, followed by 20 or more generations of brother x sister mating. If backcrossing to one parent is involved, the result is a recombinant congenic strain and must be named as such.
RILs vs near-isogenic lines (NILs)
Near-isogenic lines are the usual companion population, and the choice between them is the most common design decision. The table summarizes the Keurentjes et al. side-by-side comparison of an Arabidopsis NIL population (92 lines) and a RIL population.
| Feature | Recombinant inbred lines | Near-isogenic lines |
|---|---|---|
| Construction | Repeated selfing of segregating progeny from a two-parent cross | Repeated backcrossing so each line carries a single fragment, or a few introgression fragments, from a donor in a uniform background |
| Genome representation | Both parental genotypes represented across the panel; higher recombination than a NIL set of equal size | Mostly the recurrent parent, with small donor segments |
| Strength | Genome-wide scans; localization resolution higher than the NIL population in that comparison | Greater power to detect small-effect QTL |
| Limits | A large-effect QTL may mask detection of small additive-effect QTL; epistasis and segregation distortion can bias detection | Cannot be used to test genetic interactions; lower localisation resolution; lethality or sterility can hinder some lines |
| Replication | Population size matters more than the number of replicates | Several replicates per line are needed |
In that study, NILs detected smaller-effect QTL than RILs, although their localization resolution was lower.
RILs in QTL mapping
A quantitative trait locus (QTL) is a genomic region associated with variation in a quantitative trait. With RILs, each line is genotyped once at markers and phenotyped as often as needed; QTL are found by testing association between marker genotype and trait values across lines. Because the lines are homozygous and permanent, replication and repeated analyses of the same population are possible; the Keurentjes comparison notes that for RILs, population size is more important than replicate number.
Resolution is the main limitation. Peleman et al. (2005, PMC) note that current QTL mapping strategies generally assign a QTL to a region of 10-20 cM, and that fine mapping with near-isogenic lines requires constructing highly inbred lines over many generations. Their QTL-isogenic recombinant (QIR) method was designed to avoid repeated backcrossing; they used it to narrow a rapeseed locus (E1) to 1.1 cM.
A planning checklist: which population for which question
- Genome-wide scan for a new trait, or testing interactions between loci: RILs, since both parental genotypes are represented and NILs cannot test genetic interactions.
- A suspected small-effect locus that large-effect QTL may be hiding: NILs, which the comparison found more sensitive to small effects.
- Separating two closely linked QTL: localization resolution was higher with RILs than with NILs in the comparison; narrowing a 10-20 cM interval further needs fine-mapping material, as in the QIR approach.
- Limited phenotyping capacity: for RILs, adding lines helps more than adding replicates; for NILs, plan several replicates per line.
Reading RI strain names (MGI conventions)
For mouse and rat records, MGI designates a set with the uppercase one- or two-letter abbreviations of both parental strains, female parent first, joined by an uppercase X with no spaces. Members are numbered serially regardless of which laboratory made them.
| Name | What MGI says it is |
|---|---|
| BXD1, BXD2, BXD3 … | Members of the BXD mouse RI set, derived from C57BL/6 x DBA/2 |
| CXB | RI mouse set from BALB/c x C57BL/6J |
| HXB1, HXB2, HXB3 | Members of the HXB rat RI set, from SHR/OlaIpcv x BN-Lx/Cub |
| RIX | An F1 between two RI strains, used for mapping complex traits (for example (BXD1/Ty X AXB19/Pgn)F1) |
| Recombinant congenic | Produced when backcrossing to a parental strain is involved; not named as an RI strain |
| CC0058/Unc (Collaborative Cross) | Multiparental RI lines from eight progenitor strains, named CC, a 4-digit number, a slash and the lab code of the institution where they were bred (here Unc, University of North Carolina) |
Common misunderstandings
- “RILs and NILs are interchangeable.” They are built differently (selfing or sib-mating versus backcrossing) and answer different questions, as the table shows.
- “A RIL panel gives gene-level resolution.” Peleman et al. describe QTL mapping strategies as generally assigning a QTL to a 10-20 cM region; narrowing it further takes fine-mapping material.
- “Any inbred cross is a RIL set.” Under MGI rules, backcrossing to a parent makes a recombinant congenic strain, not an RI strain.
- “More replicates always beat more lines.” For RILs, the Keurentjes comparison reports that population size is more important than the number of replicates.
FAQ
What is the difference between recombinant inbred lines and near-isogenic lines?
RILs are fully inbred recombinants of two parents that together represent both parental genotypes; NILs are nearly identical to one parent except for one or a few introgressed donor fragments. RILs suit genome-wide scans and gave higher localization resolution in the Arabidopsis comparison; NILs detected smaller-effect QTL but cannot test genetic interactions.
How are recombinant inbred lines used in QTL mapping?
Each homozygous line is genotyped at markers and its trait measured; statistical association between marker genotype and trait across the lines locates QTL. Because the lines are permanent, the same population can be replicated and analysed repeatedly.
How many generations of inbreeding does a mouse RI strain need?
MGI describes the traditional scheme as 20 or more consecutive generations of brother x sister mating.
What does BXD mean?
BXD is the name of the mouse RI set derived from C57BL/6 (B) x DBA/2 (D), with the female parent written first; individual strains are numbered BXD1, BXD2 and so on.
Related CASRAI guides
- What is epigenetics, for heritable variation that genotype-based mapping alone does not capture
- SYBR Green real-time PCR protocol, a lab method for follow-up expression work
- Phusion PCR protocol, a general PCR amplification protocol
- Mouse body condition scoring, a general mouse welfare-scoring reference
Sources
- Keurentjes JJB et al. Development of a near-isogenic line population of Arabidopsis thaliana and comparison of mapping power with a recombinant inbred line population. Genetics. PMC1800614. Checked 2026-11-07.
- Mouse Genome Informatics. Rules and Guidelines for Nomenclature of Mouse and Rat Strains, section 4.1 Recombinant Inbred Strains and section 4.2 Collaborative Cross. Checked 2026-11-07.
- Peleman JD et al. 2005. QTL-isogenic recombinant (QIR) fine mapping. PMC1456828. Checked 2026-11-07.
