Half the cost of an in vitro plant is labor, and only some of it can be automated

Tools of the Trade

The step you automate first

Say the plan is to multiply what your lab ships and you are working out which step to spend money on first. The answer turns on where the hours actually go, and the most-cited public number on that question is nearly thirty years old.

It is a conifer review — Norway spruce (Picea abies L. Karst) carries most of the examples — but the way it splits the process into steps that can and cannot be mechanized travels further than conifers do.

Half the cost, and where that figure comes from

The review's case for automation rests on one public figure. Citing Cervelli and Senaratna (1995), it states that about 50% of the production costs for in vitro plants stem from labor, and calls that one of the few public estimates, completed already in the mid-1990s. Various private companies have estimated the economics of in vitro plant production, the review says, but those estimates are not public.

More than 20 years later, it adds, reviews still point to the high cost of labor as prohibitive for large-scale plant production in the absence of automation, citing Lelu-Walter et al. (2013). That is the whole argument for machinery: not that liquid culture is elegant, but that hands are what the plants cost. The review notes that automation is particularly attractive in countries with high labor costs, where conifer forestry is often of great economic importance.

about 50% of the production costs for in vitro plants stem from labor — Egertsdotter, Ahmad and Clapham, Frontiers in Plant Science, 2019, citing Cervelli and Senaratna (1995)

Which steps the research has managed to mechanize

The review walks the process in order — initiation, multiplication, embryo development, germination, and plant formation after transfer ex vitro — and the automation picture is different at every one.

Initiation. The review states there are no known efforts directed toward automation of this process. In conifers the usual explant is the zygotic embryo, and success requires extracting the female megagametophyte and then the embryo from within it. With the diversity of seed sizes and shapes, and the sensitivity of the target tissue, the authors write, it has not been an attractive target for automation.

Multiplication. This is where the review puts the scale-up. Cultures are maintained by repeated subculture into fresh media, with auxins and cytokinins as the common growth regulators and the nitrogen source described as important (Carlsson et al., 2017). Liquid suspension cultures of conifers can multiply at higher rates and with less aggregation than those on solid medium (Von Arnold et al., 2002), which is why liquid is preferred for this phase. Agitation breaks up the proembryogenic mass tissue into smaller aggregates that can multiply without hindrance from connected tissue (Egertsdotter and von Arnold, 1998). Liquid, the review notes, is also what enables bioreactor scale-up and automation at all.

Harvest, germination, planting. The abstract lists robotics and automation for the harvest of selected individual mature embryos, followed by automated handling of germination and subsequent planting, and says these efforts are nearing an operational stage. It does not attach a date or a manufacturer to that.

What the published record reports

None of this is a bench note from a grower; it is published work, and the review's own framing of it matters.

The opportunity to keep embryogenic cell lines for prolonged periods in cryostorage allows for field-testing before selection of production cell lines (Park et al., 2014). That is what makes the clone-by-clone economics work at all: the culture is banked while the trees are scored.

Selection takes time where trees grow slowly. In Scandinavia, the review states, trees must be at least 10 years old before selection for growth characters is meaningful, owing to poor juvenile/mature correlations, though it notes this may change with mass genomic selection.

Family forestry is the route around that wait — mass propagation of untested individuals or clones from families already tested in the breeding program. Planting with a mixture of 25 clones from six parents of tested families at each stand removes the dangers of excessive genetic uniformity associated with single clones, while exploiting the full genetic gains at the family level and avoiding the considerable expense and delay associated with effective clonal testing (Lindgren, 2008).

On whether any of this becomes an industry, the authors stay conditional: their conclusion is that with sufficient numbers of clones in production, replaced at certain intervals, SE plants can become a cost-effective, accepted and viable contribution to Swedish forestry operations provided that the required automation becomes reliably available.

For buyers

Before you take a bioreactor quote

The review's argument is that liquid multiplication is what makes bioreactor scale-up and automation possible for conifer somatic embryogenesis. Read it as a map of which steps research has mechanized, not as a shopping list: ask any supplier for throughput figures on your species and your own cell lines, and remember that on this account initiation stays a dissection job whatever else arrives on the floor.

The call

If you are costing automation for a somatic embryogenesis line, this review points you at multiplication and at the back end — harvest, germination, planting — and tells you to leave initiation to hands for now. Three caveats belong next to that.

The first is species. The biology and the economics here are conifer, mostly Norway spruce, and the review is explicit about that emphasis. A multiplication result from a spruce cell line is not a number for your Anthurium or your blueberry line.

The second is the cost figure. A 50% labor share drawn from a mid-1990s estimate is the best public number the review can point to, not a current benchmark for your lab. Your own hour counts per batch, step by step and cell line by cell line, will tell you more than any review can — the sort of thing worth logging before anyone quotes you for equipment, in xPlant or in a spreadsheet.

The third is the word nearing. The review describes automation of the later steps as nearing an operational stage as of 2019. It does not say what is on sale, and nothing in it makes a machine a substitute for the losses that still happen at maturation, germination and acclimation.