OUP user menu

Effect of acetic acid and pH on the cofermentation of glucose and xylose to ethanol by a genetically engineered strain of Saccharomyces cerevisiae

Elizabeth Casey, Miroslav Sedlak, Nancy W.Y. Ho, Nathan S. Mosier
DOI: http://dx.doi.org/10.1111/j.1567-1364.2010.00623.x 385-393 First published online: 1 June 2010


A current challenge of the cellulosic ethanol industry is the effect of inhibitors present in biomass hydrolysates. Acetic acid is an example of one such inhibitor that is released during the pretreatment of hemicellulose. This study examined the effect of acetic acid on the cofermentation of glucose and xylose under controlled pH conditions by Saccharomyces cerevisiae 424A(LNH-ST), a genetically engineered industrial yeast strain. Acetic acid concentrations of 7.5 and 15 g L−1, representing the range of concentrations expected in actual biomass hydrolysates, were tested under controlled pH conditions of 5, 5.5, and 6. The presence of acetic acid in the fermentation media led to a significant decrease in the observed maximum cell biomass concentration. Glucose- and xylose-specific consumption rates decreased as the acetic acid concentration increased, with the inhibitory effect being more severe for xylose consumption. The ethanol production rates also decreased when acetic acid was present, but ethanol metabolic yields increased under the same conditions. The results also revealed that the inhibitory effect of acetic acid could be reduced by increasing media pH, thus confirming that the undissociated form of acetic acid is the inhibitory form of the molecule.

  • Saccharomyces cerevisiae
  • xylose
  • acetic acid
  • ethanol
  • cellulose
  • inhibition


Historically, petroleum has been the major source for liquid transportation fuels. However, declining oil reserves and environmental concerns have led to interest in alternative, renewable energy sources. One promising alternative is the conversion of plant biomass into ethanol. The primary biomass feedstocks for the current ethanol industry have been corn grain and sugar cane. However, interest has recently shifted to replacing these traditional feedstocks with more abundant, non-food-based cellulosic biomass feedstocks such as agricultural wastes (e.g. corn stover) or energy crops (e.g. switchgrass). The use of cellulosic biomass as feedstock for the production of ethanol via biochemical routes presents many technical hurdles not faced with the use of corn or sugar cane as feedstock. One significant challenge is the development of efficient and economical pretreatment and enzymatic hydrolysis steps for the release of fermentable sugars from the biomass (Mosier., 2005; Stephanopoulos, 2007). Another obstacle is engineering robust, process-relevant industrial microorganisms that are capable of mixed sugar fermentation and are tolerant to inhibitors (Hahn-Hägerdal., 2007).

The complexity of cellulosic biomass hydrolysates as compared with corn starch or sugar cane hydrolysates leads to the challenge of engineering a suitable microorganism. Cellulosic biomass hydrolysates contain a variety of sugars (primarily glucose and xylose) and inhibitory compounds such as acetic acid and various phenolic compounds. The primary microorganism used in industrial fermentations, Saccharomyces cerevisiae, is not able to ferment xylose (Barnett, 1976), which represents about 40% of the sugars found in biomass hydrolysates. This problem of limited sugar utilization has been successfully addressed through metabolic engineering of S. cerevisiae with genes for the xylose metabolic pathways of xylose-fermenting bacteria, i.e. xylose isomerase (Kuyper., 2005), or yeasts, i.e. xylose reductase and xylitol dehydrogenase (Ho & Chen, 1997; Ho., 1998; Eliasson., 2000; Sonderegger., 2004). Nontraditional microorganisms for industrial ethanol production (such as Escherichia coli, Zymomonas mobilis, and Thermoanaerobacterium saccharolyticum) have also been engineered to convert mixed sugar streams to ethanol (Lindsay., 1995; Zhang., 1995; Shaw., 2008). With strains now capable of mixed sugar fermentation, the primary challenge has become the development of industrial microorganisms with tolerance to the inhibitors that are formed and released as biomass is broken down in the pretreatment and hydrolysis steps (Hahn-Hägerdal., 2007). The inhibitors found in biomass hydrolysates can be primarily classified as weak acids, furan derivatives, or phenolic compounds. These inhibitors have been shown to negatively impact the fermentative performance (cell growth, ethanol yield and productivity, and/or sugar consumption rates) of non-pentose-fermenting strains of S. cerevisiae (Palmqvist & Hahn-Hagerdal, 2000; Klinke., 2004; Almeida., 2007) and Z. mobilis (Lawford & Rousseau, 1993), as well as non-ethanol-producing strains of E. coli (Luli & Strohl, 1990). The previously studied strains are not suitable for industrial ethanol production because of their substrate and product limitations. It would be more valuable to understand how inhibitors impact microorganisms that have been engineered to ferment multiple substrates to ethanol. For example, Bellissimi. (2009) recently published a study discussing the effect of acetic acid on a strain of S. cerevisiae that was engineered for xylose fermentation with bacterial xylose isomerase. However, few inhibition studies of engineered organisms have been published.

Acetic acid is a weak acid generated from the deacetylation of hemicellulose during pretreatment (Palmqvist & Hahn-Hagerdal, 2000; Klinke., 2004; Almeida., 2007). It is known to inhibit microbial growth and has been used as an antimicrobial agent in the food and beverage industries (Luck & Jager, 1997). Acetic acid is present in varying concentrations in all types of biomass, for example corn stover and poplar contain 5.6% and 3.6% acetyl by mass, respectively (Lu., 2009). When studying the inhibitory effect that acetic acid can have on microorganisms, process-relevant conditions for the production of cellulosic ethanol at an industrial scale must be considered. Assuming that the minimum concentration of ethanol for economic distillation is 5% ethanol, the initial unhydrolyzed biomass concentration at the start of the process must be approximately 20% by weight. This would result in theoretical acetic acid concentrations of 11.2 and 7.2 g L−1 in the hydrolysates of corn stover and poplar, respectively, assuming no accumulation due to recycling of process streams. An actual acetic acid concentration of 13 g L−1 has been observed in dilute acid-pretreated corn stover hydrolysate (Lu., 2009). The removal of acetic acid and other inhibitors would add cost to the overall process. Therefore, a detailed study of the effect of these inhibitors on ethanol yields and production rates, especially for xylose fermentation, is important for ongoing microorganism development efforts and cellulosic ethanol commercialization.

The goal of the study reported in this paper was to determine the effect of acetic acid at relevant industrial process concentrations on the cofermentation of glucose and xylose under controlled pH conditions by S. cerevisiae 424A(LNH-ST), a polyploid industrial yeast strain capable of fermenting glucose and xylose. Saccharomyces cerevisiae 424A(LNH-ST) was genetically engineered for xylose metabolism by overexpressing xylose reductase, xylitol dehydrogenase from Pichia stipitis, and xylulose kinase from S. cerevisiae (Ho., 1998; Ho., 2000). We report the effects of acetic acid and pH on biomass growth, substrate consumption rates, and ethanol production rates.

Materials and methods

Yeast strain

All fermentations utilized S. cerevisiae 424A(LNH-ST), a recombinant industrial yeast strain capable of the cofermentation of glucose and xylose (Ho., 1998; Ho., 2000).

Fermentation experiments

Batch fermentations were completed in 1-L New Brunswick BioFlo 110 benchtop fermentors (Edison, NJ) equipped with pH control. The culture volume was 80% of the fermentor volume. No sparging of any gas was performed. Our measurements of dissolved oxygen during fermentation showed that the oxygen concentration declined to below 1% of saturation within 40 min of the beginning of the fermentation. The inoculum for the fermentor was prepared by pregrowing yeast aerobically in a shaker set at 28 °C and 200 r.p.m. in 2-L flasks containing 500 mL YEPD media (1% yeast extract, 2% peptone, and 2% glucose) (Mallinckrodt Chemicals, Phillipsburg, NJ).

YEP media (1% yeast extract, 2% peptone) were used as the fermentation media. YEP was chosen as the fermentation medium because the extensive prior data for the performance of this yeast were obtained using this medium. This allowed for comparisons between this work and prior work. In addition, YEP is a rich medium. Thus, the most significant stress on the fermenting yeast was the presence of acetic acid unencumbered by the metabolic stress due to the synthesis of minor metabolites. Glucose and xylose concentrations in the starting fermentation media were 60 g L−1 each. The acetic acid concentrations examined were 0 (for control), 7.5, and 15 g L−1. These concentrations were selected to represent the range of concentrations that can be expected in hydrolysates from a variety of biomass sources (Takahashi., 1999). The pH of the media was adjusted to the desired value (5, 5.5, or 6) with ammonium hydroxide (28–30% NH3) (Mallinckrodt Chemicals) or 14.8 M phosphoric acid (Fisher Scientific, Pittsburgh, PA).

Once the cell density of the growth culture reached approximately 6 g dry cells L−1, the culture was centrifuged for 5 min at 3100 g. The cell pellet was resuspended in YEP and used to inoculate the fermentor to an initial cell density of 4.75 g dry cells L−1. The temperature and r.p.m. for the fermentation were set to 28 °C and 200 r.p.m., respectively. The media pH was continuously controlled within ± 0.1 from the desired value using the pH control system provided with the BioFlo 110 fermentor using 1 M phosphoric acid and 1 M ammonium hydroxide. biocommand plus software (New Brunswick Scientific Co.) was used to record the real-time media pH. The fermentation then proceeded for a maximum of 240 h and all fermentations were carried out in duplicate.

Analysis of fermentation substrates and products

The fermentation metabolites were analyzed by HPLC using the method outlined by Lu. (2009) using a Waters Alliance 2695 HPLC system with an Aminex® HPX-87H 300 × 7.8 mm column (Bio-Rad Laboratories, Hercules, CA). The HPLC column operating conditions were 60 °C at a flow rate of 0.6 mL min−1 for the mobile phase, 5 mM sulfuric acid in water.


We were interested in determining the combined effects of acetic acid and pH on various fermentation performance characteristics of S. cerevisiae 424A(LNH-ST). Specifically, we examined the impact of these factors on biomass growth, glucose and xylose consumption rates, and ethanol productivity rates and yields. To accomplish this, a 32 factorial experimental fermentation design was selected. Two factors, acetic acid concentration and media pH, at three different levels were chosen to provide a total of nine different fermentation conditions, with each fermentation condition repeated twice. Figure 1 shows the representative fermentation profiles for each of the nine conditions. A visual comparison of the profiles revealed the effect of acetic acid and pH on the cofermentation of glucose and xylose. The maximum biomass concentrations decreased and the xylose consumption rates slowed as the acetic acid concentration increased and pH decreased. The increase in the acetic acid concentration and the decrease in pH also corresponded to a decrease in ethanol production rates. Acetic acid had a minimal inhibitory effect on glucose consumption, with the exception of the most severe condition tested (pH 5 and 15 g L−1 acetic acid). More detailed results for each of these fermentation performance characteristics are provided in the following sections.

Fig. 1

Time course profiles for the cofermentation of glucose and xylose by Saccharomyces cerevisiae 424A(LNH-ST) in the presence of varying concentrations of acetic acid and pH values: (a) 0 g L−1 acetic acid, pH 6; (b) 0 g L−1 acetic acid, pH 5.5; (c) 0 g L−1 acetic acid, pH 5; (d) 7.5 g L−1 acetic acid, pH 6; (e) 7.5 g L−1 acetic acid, pH 5.5; (f) 7.5 g L−1 acetic acid, pH 5; (g) 15 g L−1 acetic acid, pH 6; (h) 15 g L−1 acetic acid, pH 5.5; and (i) 15 g L−1 acetic acid, pH 5. •, glucose; ○, xylose; ▾, acetic acid; ▵, ethanol; ♦, biomass.

Impact of acetic acid and pH on biomass growth

Because of the high cell concentration at inoculation, minimal cell growth was observed. However, the extent of this growth for each of the nine fermentation conditions was compared to determine the effect of acetic acid and pH on the growth of S. cerevisiae 424A(LNH-ST). The maximum biomass concentrations under each condition are summarized in Table 1. At a given pH, the maximum biomass concentration was shown to decrease significantly in fermentations with acetic acid as compared with the control (with the exception of the least severe condition tested, pH 6 and 7.5 g L−1 acetic acid).

View this table:
Table 1

Summary of biomass and substrate consumption results under each fermentation condition for the cofermentation of glucose and xylose by S. cerevisiae 424A(LNH-ST)

pH 6pH 5.5pH 5
Total acetate+acetic acid concentration (g L−1)07.51507.51507.515
Maximum biomass concentration (g L−1)6.962AB± 0.0606.545ABC± 0.3576.129CD± 0.1787.14A± 0.006.307BCD± 0.006.158CD± 0.0897.051A± 0.0306.218CD± 0.0305.772D± 0.060
Initial specific glucose consumption rate (g glucose g−1 dry cell h−1)1.867A± 0.0091.651ABC± 0.1111.610BC± 0.0341.729AB± 0.01991.571C± 0.0621.198D± 0.0391.820AB± 0.0190.975D± 0.0200.217E± 0.001
Initial specific xylose consumption rate (g xylose g−1 dry cell h−1)0.354A± 0.0090.201C± 0.0010.123D± 0.0020.285B± 0.0020.138D± 0.0010.045E± 0.0020.272B± 0.0010.032E± 0.0020.014F± 0.003
  • Values listed are the mean and SE of two duplicate fermentations for each condition. Means with the same letter are not significantly different at a confidence level of 90%.

Slight reductions in the biomass concentration were also observed as the pH decreased for fermentations at a given acetic acid concentration. This effect was not observed in the control fermentations, suggesting that the reductions in biomass when acetic acid was present cannot be explained solely by the decrease in pH. This hypothesis was confirmed by a previous study that reported that pH had no effect on biomass growth in the range examined here (Phowchinda., 1995).

The biomass yields observed (0.016–0.025 g g−1 sugar consumed) were less than the expected yield of 0.055 g g−1 sugar consumed for anaerobic metabolism (Davis., 2006). A carbon balance of all fermentations following the convention of Wang. (1979) closed at or above 90% (data not shown). Therefore, the lower yield of biomass was offset by the higher yield of fermentation products such as ethanol, CO2, xylitol, and glycerol.

Impact of acetic acid and pH on specific glucose and xylose consumption rates

To explore the effect of acetic acid and pH on substrate consumption, the initial specific sugar consumption rates (both glucose and xylose) were calculated for each fermentation condition. This rate was calculated by determining the slope of the steepest portion of the substrate concentration curve and dividing that by the average cell concentration during that period to yield a rate with the units of g substrate g−1 dry cell h−1. The initial specific glucose consumption rates are summarized in Table 1. The glucose consumption rates for a single pH condition decreased significantly with increasing acetic acid concentration, with the exception of pH 6. Under the most severe condition (pH 5 and 15 g L−1 acetic acid), the glucose consumption rate was only 12% that of the control. However, the inhibitory effect of acetic acid on the glucose consumption rates decreased as pH increased. For example, the glucose consumption rate of the fermentation with 7.5 g L−1 acetic acid at pH 5 was about half that of the control, compared with almost 90% of the control when the pH was increased to 6.

The initial specific xylose consumption rates are also summarized in Table 1. It is evident that xylose consumption was strongly inhibited by acetic acid; a significant decrease in the rate was observed with both increasing acetic acid concentrations and decreasing pH values. Under the harshest condition (pH 5 and 15 g L−1 acetic acid), no significant xylose consumption was observed. Results from a similar study (Bellissimi., 2009) did not show such a severe decrease in the xylose consumption rate for the fermentation conditions of pH 5 with 3 g L−1 acetic acid. However, the concentrations examined in this study were 2.5–5 times greater than those examined by Bellissimi and colleagues.

To better understand the source of the inhibition shown in Table 1, the equilibrium concentration of acetic acid was calculated for each fermentation. The calculation used the Henderson–Hasselbach equation to determine the acetic acid–acetate equilibrium in the medium using a pKa of 4.75 for acetic acid and the pH for each fermentation condition. The initial specific xylose consumption rate was plotted vs. the calculated undissociated acetic acid concentration in Fig. 2. This shows a strong correlation between the data and an exponential decay function in the xylose consumption rate as the undissociated acid concentration increases (R2 of 0.95). Similar studies have also observed an exponential decay in S. cerevisiae-specific growth rate and fermentation rate for glucose fermentations with increasing acetic acid concentration (Pampulha & Loureiro, 1989; Narendranath., 2001).

Fig. 2

Relationship between the initial specific xylose consumption rate and the undissociated acetic acid concentration.

Impact of acetic acid and pH on ethanol production

To investigate the effect of acetic acid and pH on ethanol production, ethanol metabolic yields and ethanol volumetric production rates were calculated. The ethanol metabolic yields, calculated by dividing the observed ethanol concentrations by the theoretical ethanol concentrations (0.51 times the total concentration of consumed glucose and xylose), are provided in Table 2. If all of the consumed glucose and xylose was converted to ethanol, the ethanol metabolic yields would be equal to 1. However, a portion of the consumed sugars resulted in the production of glycerol, xylitol, and/or additional cell mass. Thus, the metabolic yields are all <1. The presence of acetic acid enhanced rather than inhibited the yield of consumed sugars to ethanol, as seen by an increase in the ethanol metabolic yields as the concentration of acetic acid increased. At pH 5, ethanol metabolic yield improvements of 20% and 30% were observed in the presence of 7.5 and 15 g L−1 acetic acid, respectively, as compared with the control.

View this table:
Table 2

Summary of ethanol yield and productivity results under each fermentation condition for the cofermentation of glucose and xylose by Saccharomyces cerevisiae 424A(LNH-ST)

pH 6pH 5.5pH 5
Total acetate+acetic acid concentration (g L−1)07.51507.51507.5 g L−115 g L−1
Ethanol metabolic yield0.785BC± 0.0120.857ABC± 0.0020.875AB± 0.0080.793BC± 0.0060.826BC± 0.0120.880AB± 0.0080.755C± 0.0020.892AB± 0.0620.956A± 0.018
Ethanol volumetric production rate (g ethanol L−1 h−1)1.234A± 0.0270.784C± 0.0370.588D± 0.0091.177A± 0.0280.574D± 0.0440.474DE± 0.0130.941B± 0.0060.453E± 0.0080.403E± 0.006
  • Values listed are the mean and SE of two duplicate fermentations for each condition. Means with the same letter are not significantly different at a confidence level of 90%.

In addition, ethanol volumetric production rates (Table 2) were calculated by dividing the maximum ethanol concentration by the fermentation time required to reach that concentration. Significant decreases in productivity were seen in the presence of acetic acid. Comparing the least and the most severe conditions (0 g L−1 acetic acid at pH 6 and 15 g L−1 at pH 5, respectively), a 67% decrease in volumetric productivity was observed. The results also show a linear increase in the production rates with increasing pH (R2 of 0.89 and 0.99 for 7.5 and 15 g L−1 acetic acid concentrations, respectively, plots not shown).


The effect of acetic acid on non-pentose-fermenting S. cerevisiae strains has been studied widely (Maiorella., 1983; Pampulha & Loureirodias, 1989; Phowchinda., 1995; Taherzadeh., 1997; Thomas., 2002). Prior studies examined a range of acetic acid concentrations (up to 12 g L−1) under a variety of uncontrolled or controlled pH values (pH 2.8–5.5). A common observation was that acetic acid resulted in a decrease in biomass yield coupled to an increase in ethanol yield (Maiorella., 1983; Taherzadeh., 1997; Thomas., 2002). However, the increased ethanol yields came at the cost of a decreased ethanol production rate (Phowchinda., 1995). The decrease in the fermentation rate was explained by a decrease in intracellular pH, leading to the conclusion that it is the concentration of the undissociated, uncharged form of acetic acid that governs the inhibitory effect (Pampulha & Loureirodias, 1989). Undissociated acetic acid freely diffuses across the cell membrane and rapidly dissociates because of the higher intracellular pH, resulting in the release of protons into the cytoplasm. Plasma ATPase pumps these protons out of the cell at the cost of ATP to avoid intracellular acidification until the influx of protons exceeds the cell's proton-pumping capability and acidification of the cytoplasm cannot be avoided (Russell, 1992). ATP is also needed for the removal of excess acetate from the cytoplasm through the energy-dependent weak acid efflux pump, Pdr12 (Piper., 1998). The need for ATP to maintain intracellular pH homeostasis and reduce the internal acetate concentration explains the decreased biomass and increased ethanol yields; glucose was diverted from biomass generation to ethanol production to generate the ATP needed for cell maintenance. Similar to the studies discussed above, a decreased biomass concentration and increased metabolic ethanol yields were seen in the glucose/xylose cofermentations with acetic acid for the given study. These results can also be explained by the diversion of carbon from biomass growth to ethanol production for ATP generation. The ethanol production rates also decreased in the presence of acetic acid in this study. Upon comparison of the results, the rate Phowchinda. (1995) achieved with 6 g L−1 acetic acid and uncontrolled pH conditions was approximately the rate we achieved under the harshest condition tested in the present work (15 g L−1 at pH 5). This suggests that S. cerevisiae 424A(LNH-ST) was more tolerant to acetic acid than the strain they used and/or controlling the media pH can significantly mitigate inhibition by acetic acid.

Although there were many similarities in the experimental procedures and results from this study and the previous studies, the difference in microorganisms (pentose vs. non-pentose-fermenting strains) makes a direct comparison difficult. A limited number of studies have been published that investigate the impact of acetic acid on genetically engineered S. cerevisiae strains capable of both glucose and xylose fermentation (Helle., 2003; Bellissimi., 2009). Bellissimi. (2009) looked at the effect of acetic acid on an S. cerevisiae strain engineered with the xylose isomerase pathway as opposed to the xylose reductase and xylitol dehydrogenase pathway in the strain used in this study. The conditions they examined were an acetic acid concentration of 3 g L−1 under controlled pH values of 3.5 or 5. Helle. (2003) studied the effect of acetic acid on an S. cerevisiae strain that was engineered with the xylose reductase and xylitol dehydrogenase genes for xylose metabolism, similar to the strain used in this study. Their experimental conditions were 3 g L−1 acetic acid and initial pH values of 4, 4.7, and 5.5. Similar to the results with the non-pentose-fermenting S. cerevisiae strains, biomass yields decreased in the presence of acetic acid for both studies. Bellissimi. (2009) also showed that xylose consumption rates were more affected by acetic acid than glucose consumption rates, while Helle. (2003) noted a decrease in ethanol production rates from the presence of acetic acid in the fermentation media. Similar observations were made in the present study, confirming the impact of acetic acid on the cofermentation of glucose and xylose. The primary difference from these results, when compared with the results using non-pentose-fermenting S. cerevisiae strains, was the increased inhibitory effect on xylose consumption vs. glucose consumption. The rate of xylose consumption in S. cerevisiae 424A(LNH-ST) is approximately 20% that of glucose and the ATP yield is less (1.67 mol ATP mol−1 xylose compared with 2.0 mol ATP mol−1 glucose). Thus, the estimated ATP generation rate when xylose is the sole carbon source is approximately 17% of the ATP generation rate when glucose is fermented. This is likely a major reason why the inhibitory effect of acetic acid on xylose consumption is more severe than on glucose consumption. Bellissimi and colleagues showed that a continuous feed of glucose at low concentrations releases some of the inhibition of xylose consumption, further supporting this conclusion. Bellissimi. (2009) also observed increasing glucose consumption rates with 3 g L−1 acetic acid present, a trend not seen in this study. Low amounts of acetic acid have been shown to stimulate fermentation (Taherzadeh., 1997; Thomas., 2002). However, the acetic acid concentrations tested in this study exceeded those amounts by a factor of 2 or more.


The presence of inhibitory compounds in biomass hydrolysates is a major obstacle facing the cellulosic ethanol industry because they impact the fermentative performance of microorganisms negatively. The effect of acetic acid, one of these inhibitors, on the cofermentation of glucose and xylose by S. cerevisiae 424A(LNH-ST) under controlled pH conditions was examined. Acetic acid inhibited biomass growth, substrate consumption, and ethanol volumetric productivity. However, acetic acid enhanced ethanol metabolic yield. Significant decreases in the maximum biomass concentration were observed in the presence of acetic acid, with this effect becoming less severe as the pH was increased. Similar trends were noted with glucose and xylose initial specific consumption rates, with the effect being more significant on xylose consumption. An exponential relationship was found between the initial specific xylose consumption rates and the concentration of undissociated acetic acid, confirming that the inhibitory effect of acetic acid is linked to the undissociated form of acetic acid. Ethanol production rates decreased considerably in the presence of acetic acid. However, no inhibitory effect was seen with the ethanol metabolic yields; rather, acetic acid was shown to improve these yields. The impact of media pH was also investigated in this study. The results suggest that increasing media pH can alleviate some of the inhibitory effect of the acetic acid as this causes a decrease in the concentration of undissociated acetic acid, the inhibitory form of acetic acid for S. cerevisiae fermentations.


This work was supported by the US Department of Energy Biomass Program, Contract GO17059-16649, and Purdue Agricultural Research Programs. This material is based on work supported under a National Science Foundation Graduate Research Fellowship. Any opinions, findings, conclusions, or recommendations expressed in this publication are those of the authors and do not necessarily reflect the views of the National Science Foundation. The authors thank Haroon Mohammad for the preliminary work he completed on this project. They also thank Eduardo Ximenes and Chialing Wu for their internal review of this manuscript.


  • Editor: Jens Nielsen


View Abstract