Tomato Breeding 101

The cultivated tomato is a self-pollinated plant.   With few exceptions, the stigma (female) remains within the anther cone (male) until after pollination, greatly limiting potential cross-pollination.  It is estimated that >95% of the seed in tomatoes is normally the result of self-pollination.

Multiple generations of selfing inevitably result in a single allele being fixed at each genetic locus - each gene/locus is homozygous for a particular allele.  Inbred lines are stable from generation to generation (i.e true breeding).  Thus the progeny from saved seed from an inbred line or OP (open pollinated) tomato variety will look just like the female parent.  A progeny plant not identical to the female parent is likely the result of chance cross pollination, facilitated by one of various pollinating bees frequenting most gardens.

There is no genetic variability within a stable inbred line.  All plants generated from such a line are genetically identical to one and other.  No genetic variation, no potential for breeding/improvement … its that simple.  A breeding population with genetic variability must be derived from a cross/crosses between inbred lines.  Here are a few examples:

1)   OP (selfed) seed is harvested from a commercial F1 tomato hybrid (e.g. SunGold F1).  The F2 progeny will segregate for all traits that were different between the two parents.  Usually the identity of the parents is proprietary to the breeder/seed company.  Several generations of selfing/selection may create an OP population that looks something like the F1 hybrid – but no guarantee.

2)   A plant tracing to a chance outcross is identified in OP progeny from an heirloom/OP variety (e.g. Brandywine).  Again the F2 will segregate.  In this example you know the female, but not the male parent.  There is a high degree of luck required to pull something useful out of a chance outcross.  However there are some sterling examples of success: Lucky Cross, Berkeley Tie Dye, Earl’s Faux ….

Fruit from a couple of 2010 hand crosses

3)   A breeder plans a cross between two parents that compliment each other in some way - for example an early, small fruited, high yielding and disease resistant line crossed with a late, large fruited, lower yielding, and very tasty variety (e.g. SunSugar x Lucky Cross).  From the segregating F2 – F6 generation progeny the breeder selects the plants that best combine the desirable traits from each parent.

A F1 cross between two inbred parents is very uniform, and no selection is possible.  The maximum genetic variation is expressed in the F2 generation, and it is important to have as many F2 progeny as possible to maximize the opportunity to capture the desired characteristics from each of the two parents.  With each succeeding generation there is less genetic variation, and population size can be smaller.  By the F6 generation selected lines should be pretty stable and true breeding. A F1 cross between non-inbred parents, for example the double cross (AxB) X (CxD) or three way cross (AxB) x C, will segregate in the F1 generation, so selection within multiple F1 progeny is often desired - subsequent filial generations can be handled as described above.

Fruit from 3 segregating F2 progeny (SunSugar x Pink Boar)

Segregating for epidermis color in the F2

Segregating F1 from a three-way cross

Some important traits are controlled by many genes (e.g. yield and taste).  These are called quantitative traits and heritability is complex.  Many important traits in tomato are controlled by single genes.  A good summary of common tomato genes and their alleles can be found here.  A recessive allele is only expressed in the homozygous condition (e.g. gs/gs = homozygous for green stripe).  A plant showing a recessive phenotype will be true breeding for that trait in subsequent generations.   A homozygous dominant can only be indentified by progeny testing (i.e. evaluating the phenotype of numerous progeny).  When all the progeny have the dominant phenotype, the parent can be assumed to be homozygous for the dominant allele.

A successful tomato breeding program will generally require careful selection of parents for new crosses, evaluation of enough F2 plants to find the desired combination of traits from the parents, and active observation/notes and selection in each of the segregating generations.   Our experience is that few crosses will yield an end product that is truly unique and special; plant breeding is both an art and a science; and there’s an element of luck and unpredictability that make it fun.

One final note – this is something everyone can do in their own garden, no matter how small.  Create something new, involve your kids, teach them a little about genetics and plant science.  If you are looking for starting materials, or want to learn about the experience of others, check out one or both of two great tomato on-line forums: Tomatoville and The Tomato Depot.

2014 update:  the above breeding outline describes a typical breeding program for O.P. (open pollinated) varieties.  A breeding program for F1 hybrids would include all of the above elements plus evaluating O.P./inbred lines for suitability as a potential parent in one or more new F1 hybrids.  This final step will normally require many dozen test crosses, evaluation of lines for parent traits (i.e. suitability for being a female or male parent), and careful consideration of combining parents with complimentary attributes.  As discussed in a recent post on this site (link), F1 hybrids increase complexity in both breeding and seed production, but have intrinsic advantages vs. O.P varieties in many situations.