Using a Taguchi DOE to investigate factors and interactions affecting germination in Miscanthus sinensis

The Miscanthus genus of perennial grasses is grown for bioenergy and biorenewable feedstocks. Most Miscanthus crop is M × giganteus which is rhizome propagated and therefore difficult to multiply at large scale. Seed-based propagation of new hybrids is being developed, but Miscanthus is difficult to establish from seed especially in the field. Miscanthus is often grown on marginal land adding to the challenge of successfully establishing the crop. Improved understanding of the limits and biology of germination in Miscanthus species is needed. Seed germination is affected by physical and chemical factors that impact germination differently depending on level of exposure. In this investigation of Miscanthus germination, four hormones plus water stress were investigated and the range over which these factors affect germination was determined. An efficient Taguchi experimental design was used to assess the five factors in combination with the effects of light and seed priming. This determined an example of a set of optimum conditions for Miscanthus germination and demonstrated how this could change based on fixing one condition. The experiment showed how environmental stress impacted germination and how treatments such as gibberellic acid could be used to mitigate stress.


Water stress (PEG)
The PEG range finding experiment was designed to align with the NaCl experiment in terms of water potential (Ψ). PEG avoided any effect of ion toxicity associated with NaCl; however, other problems are possible for example, large PEG molecule could hypothetically become immobilised within the blue germination paper and there is some evidence that low molecular mass PEG molecules could penetrate a plant or seed and cause a toxicity 17,18 . As such it is unclear if high or low Molecular weight PEG would be preferable and so both PEG 4000 and 8000 were tested. To compare and align osmotic pressures different equations were requried for NaCl and PEG. NaCl obeys the Hoff Equation (1) but PEG which is a large coiled molecule binds water according to the Tyndall effect 19,20 . 21 used an equation (2) to model the osmotic potentials of different non-ideal solutions. α and β are listed for a variety of PEG molecular masses, from these a variety of curves can be calculated of the water potential in solutions of PEG. The water potentials using PEG 8,000 and 4,000 were calculated using the best suited equation. The general PEG 8,000 equation was used (3) from 22 ; where T is temperature in degrees Celsius and Π is the water potential in bar. The 22 equation (3) would be best for calculating a very wide range of PEG solutions; however, the PEG 8,000 specific calculation (2) 21 was used because there is a large difference between PEG 6,000 and PEG 10,000 22 .
Priming Seed was water primed by imbibing until the seed were close to germination then drying back to the original moisture content. The priming process was performed by Elsoms Seeds (Spalding, UK), the same seed are used in 23 The effects of priming on germination and early growth of Miscanthus seed were tested in two experiments. Primed and unprimed M. sinensis seed were germinated on two plates, 64 seed of each, and measurements taken daily. Germination data were analysed using a Kruskal-Wallis rank sum or a t-test depending on the normality of the data. A second replicated experiment into priming used three dishes separated into two sections: 50 primed and 50 unprimed M. sinensis seed were randomly assigned to left or right side. The seeds were grown for 37 days to allow more difference in measurements of epicotyl and root growth. Fluorescence images for each plate were gathered through-out the experiment and at the end of the experiment using the CF Imager (Technologica Ltd, Colchester UK). Water was added as necessary, to keep the blue germination paper fully wet. The mean fluorescence Fv/Fm values and mean total fluorescence areas for each side of the dish were analysed with a paired t-test or a Wilcoxon signed rank when lacking a normal distribution. Epicotyl and root elongation was analysed via t-tests or Kruskal-Wallis rank sums depending on normality. A Kruskal-Wallis was also used to determine if there was a significant difference in the total germination counts at the end of the experiment.

Abscisic acid
The final mean elongation of the epicotyls exposed to concentrations of GA that were more than or equal to 0.75 mg L −1 was 22.1 mm, which is longer than the control at 7.7 mm (Fig 1). Root elongations did not appear to be affected by the GA concentration. The time taken for the individual seeds to germinate under the influence of GA showed no clear pattern. GA concentration produced maximum germination between 1 and 10 mg L −1 This is most noticeable for GI, which was 4.09 in the control, rising to 4.65 at 7.5 mg L −1 and between 3.3 and 4.4 in the remaining treatments.

Gibberellic acid
There was a negative effect of ABA on root and epicotyl elongation (Fig 1), particularly at higher concentrations (> 20 mg L −1 ). However, this is only a difference between a median of 5 mm for the control and 7 mm for concentrations less than or equal to 5 mg L −1 . The mean differences for both is around 7.19 mm. Time to 50% germination was 4 days for the control and for 0.05 mg L −1 , before increasing to 8 days at 0.1 mg L −1 after which 50% of seeds do not germinate. GI dropped from over 4 in the first two concentrations to 3.6 at 0.1 mg L −1 then 2.5 at 0.5 mg L −1 . It then stabilised around 2 mg L −1 until more than 30 mg L −1 after which it remained less than 1.

Brassinosteroid
Brassinosteroid (BR) had no effect on epicotyl elongation and only a minor effect on root elongation (Fig 1), which decreased at concentrations above 0.5 mg L −1 , and both epicotyl and root elongation fluctuated more above 0.1 mg L −1 . The change in mean root elongation occurred from a mean of 6.6 mm at 0 mg L −1 to a mean of 2.5 mm as an average of the highest three concentrations (1, 1.5 & 2 mg L −1 ). Epicotyl elongation went from 7.4 mm to 4.5 mm over the same range of concentrations. The median time taken for seeds to germinate remained at two days at concentrations below 0.5 mg L −1 . At concentrations of 0.5 or 1 mg L −1 , the median time to germinate was three days and four days at higher concentrations. Only 22% of seed germinate at 1.5 mg L −1 ; however, this result appears to be an outlier, because across all other dishes germination was between 41% and 53% regardless of BR concentration.

Auxin
Auxin (NAA) appeared to have a positive effect on epicotyl and root elongation at 0.01 to 0.05 mg L −1 , after which the root elongation was less than the control (Fig 1). Epicotyl elongations was unaffected or greater until over 50 mg L −1 auxin. It was observed during the test that the roots appeared fluffier, probably due to more root hairs being visible in the auxin dishes. Auxin had little noticeable effect on the speed of germination. The mean number of days for a seed to germinate stayed at around 2.5 days regardless of concentration. However there was a small increase in median germination speed from 2.5 to 2 days comparing the control treatment to all concentrations of auxin.

Sodium chloride
Germination time increased below -0.1 MPa (20 mM) (Fig 1). This was a small effect and larger changes in germination time were not seen until water potential reached less than -1 MPa (201 mM). GI showed similar patten as the proportion of seed germinated and at the end of the test dropped approximately 12.8% per MPa. Little germination was recorded in the higher salt concentrations, at -2.2 and -4.1 MPa (444-827 mM). Epicotyl elongation decreased linearly with the nonlinear x-axis (Ψ). At water potentials lower than -0.8 MPa (161 mM) the increase in salt had a positive effect on root elongation, as the pressure increased the effect was negative similar to the effect of salt on epicotyls.

Water stress (PEG)
Epicotyl elongations and to a lesser extent root elongation declined with water potential (Ψ) in PEG treatments (Fig 1). The decline began at 0.05 MPa in both epicotyl and root but decreased more rapidly after a Ψ of 0.3 MPa for epicotyl elongation and after a Ψ of 1.2 MPa for root elongation. In both cases, seeds were more affected by the PEG 8000 than the PEG 4000, as seen in the more consistent drop for PEG 8000. There was a sharper decrease of seedling epicotyl elongation in PEG 8000. Time taken for seeds to germinate rose sharply around -0.5 MPa for both PEG 4000 and 8000. Seed in PEG 4000 germinated more slowly and over an inter-quartile range of 2 to 6 days, compared to PEG 8000, where germination occurred over an inter-quartile range of 2 to 3 days. Seeds at levels of -1.6 and -2.2 MPa had a germination proportion of ∼0.1 in PEG 4000 and ∼0 in PEG 8000.

Priming
Epicotyl elongation was the only significant response to priming in the first priming test; control seed produced longer epicotyls at 17.1 mm compared to 13.3 mm from primed seed. There was no significant difference in germination percentage or epicotyl 3/8 length in the second priming experiment but the priming treatment resulted in seedlings with a significantly shorter mean root length according to a comparison of means using Student's t-test. The mean dark-adapted chlorophyll fluorescence response (Fv/Fm) was significantly higher in primed seed when tested using a Wilcoxon signed rank (P < 0.01). However, the mean total fluorescence area was significantly higher in unprimed seedlings (P < 0.01).