1 Forest Genomics Research and Development in Canada: Priorities for 1 Developing an Economic Framework 2 Ilga Porth (porth@mail.ubc.ca)1, Mark Boyland (Mark.Boyland@NRCan-RNCan.gc.ca)2, 3 Suborna Ahmed (suborna@alumni.ubc.ca)2, Yousry A. El-Kassaby (y.el-kassaby@ubc.ca)1,2 and 4 Gary Bull (gary.bull@ubc.ca)2* 5 Address: 1Department of Forest and Conservation Sciences, University of British Columbia, 6 2424 Main Mall, V6T 1Z4 Vancouver, BC, Canada; 2Department of Forest Resources 7 Management, University of British Columbia, 2424 Main Mall, V6T 1Z4 Vancouver, BC, 8 Canada 9 *Correspondence: Gary Bull 10 Faculty of Forestry, University of British Columbia 11 Vancouver, British Columbia, V6T 1Z4 12 CANADA 13 Tel. 604 822 1553 14 15 ABSTRACT 16 Forest genomics is a relatively recent research field, and is often poorly understood both by the 17 public and forest managers. Genomics in forestry, an expansion of forest biotechnology, seeks to 18 develop generalized technologies for use in industrial plantations and/or natural forests as well as 19 within process optimization, product development and international trade facilitation. With such 20 tools it is possible to address formerly intractable issues, such as understanding the 21 underpinnings of complex traits for conservation management purposes, improved use of forest 22 trees as carbon sinks, feedstock for biofuels and “green chemistry” through deeper understanding 23 2 and effective utilization of forests’ natural variation. Diverse end-users could benefit from 24 genomics tools, for example, real-time detection and mapping of known and novel pathogens 25 along with risk assessments to protect forest nurseries and natural forests from invasive 26 pathogens and reduce economic losses associated with diseases. Since 2001, there has been 27 approximately $123 million investment in Canadian forest genomics research and we thought it 28 would be helpful to summarize the various projects in Canada and the USA and identify the 29 research priorities and potential economic implications, by (a) developing a robust typology of 30 forest sector genomics research relevant to Canadian application, (b) categorizing each initiative 31 for its application potential (commercial; non-commercial), and (c) demonstrating with 32 silvicultural gain, insect resistance, and wood composition themes the application of modeling 33 and economic analysis. 34 35 Keywords: Forest genomics, research projects inventory, thematic areas, commercial 36 applications, social and ecological dimensions, economic analysis, pest resistance, silvicultural 37 gain, wood composition, market access, tree improvement, tools 38 3 Introduction 39 Forests have important economic, ecological and social values. Forest trees contribute 40 substantially to global carbon uptake, are an integral part of complex ecosystems, and provide 41 benefits for human wellbeing in the health and recreational sector. Canadian forests are 42 important for the country economy [the forest sector contributed $19.8 billion to the Canadian 43 gross domestic product (GDP) in 2013] but also for global carbon cycle as Canadian forests 44 constitute roughly 10% of the earth’s forests (396,000,000 hectares) (Aukema et al. 2009; Rank 45 and Associates 2013). Therefore, healthy Canadian forests are of paramount importance locally 46 and globally. Through the past decade, the Canadian forest sector has faced strong structural and 47 deep cyclical challenges: low cost competition; digital media; reduced demand through global 48 economic downturn. In addition, major insect devastations in both Eastern and Western Canada 49 by spruce budworm, emerald ash borer, and the mountain pine beetle, respectively, affecting 50 28,000,000 hectares of forest land between the years 2009 - 2010 alone provided enormous 51 environmental challenges. These challenges have combined to result in a 30% reduction to the 52 sector’s contribution to GDP since 2007 (Rank and Associates 2013) and over 130,000 job losses 53 (Kumagai et al. 2010). 54 Canadian forests are mostly publically owned and therefore are subject to public forest 55 policy with its regulations, legislation and directives. Another characteristic of Canadian forests 56 is slow growth resulting in long rotation cycles of 80 years or longer, yet the rotation cycles are 57 strongly species dependent. Hence, an improvement in selection gains for productivity, wood 58 quality and other value-added characteristics, climate adaptability, insect and pathogen resistance 59 along with shortening tree breeding cycles by genomics-informed early selection of superior 60 genotypes with such advantageous trait characteristics would significantly increase forest 61 4 productivity (Costanza and McCord 2009). Therefore, it is worthwhile to undertake efforts to 62 drastically shorten the breeding cycles by implementing DNA fingerprinting methods to achieve 63 the anticipated genetic gain in tree breeding without the need to establish controlled crosses 64 and/or progeny trials (El-Kassaby and Lstiburek 2009; Resende et al. 2012a). 65 The past decade has seen approximately $123 million investment in forest genomics 66 research (summarized in the current study, see Supplementary Material), and investment in forest 67 genomics in Canada started in the year 1999 (Kumagai et al. 2010). However, this relatively 68 recent research field, forest genomics, is often poorly understood by the public, including forest 69 managers and is often confused with genetic engineering (GE). This misconception is attributed 70 to poor communication and the dire lack of providing clear distinction between these two areas 71 of research. Genomics technologies in forestry represent an expansion of forest biotechnology. 72 Genomics is defined as the branch of biotechnology that deals with the DNA sequence of the 73 entire organism’s genome (chromosome set) and includes two major disciplines; namely, 74 functional (determines the biological function of all genes and their products) and structural 75 (determines the three-dimensional structures of proteins). Genomics’ purpose in tree breeding 76 technologies (Bhalerao et al. 2003; Costanza and McCord 2009) is foremost to describe and 77 make efficient use of the abundant natural genetic variation present within undomesticated forest 78 tree populations (Groover 2007) and contribute to an understanding of the genetic architecture 79 underlying (ecological or industrial) traits of interest on a genome-wide scale (Sederoff et al. 80 2009; Street et al. 2006; Holliday et al. 2008; Porth et al. 2012; Prunier et al. 2013; Verta et al. 81 2013; McKown et al. 2014; Kirst et al. 2004; Porth et al. 2013a; Beaulieu et al. 2014). While GE 82 is the deliberate alteration of the genetic material of living organisms involving the introduction 83 of alien genes from other taxa resulting in the production of recombinant DNA which in turn is 84 5 used to create products (proteins) the original organism is incapable of producing without this 85 alteration (Porth and El-Kassaby 2014). Thus, a distinction between the “improvement” achieved 86 through GE and that of classical tree improvement (i.e., traditional tree breeding) is of great 87 importance as the former deals with introduced alien genetic material while the latter capitalizes 88 on species’ natural variation for capturing gain. In addition, the implementation of genomics 89 research, i.e. its economic potential for forestry, can be relatively wide-ranging, encompassing 90 tree breeding and selection to non-breeding opportunities. 91 Yet, forest genomics can also develop generalized technologies that may be applied to 92 industrial plantations as well as to natural forests. With such tools in hand it may be possible to 93 address new or formerly intractable issues, such as the understanding of the genetic underpinning 94 of complex traits for applications in conservation and management of natural forests, the 95 improved use of forest trees as carbon sinks, and the use of forest trees as biofuels feedstock and 96 in contributions to “green chemistry” (renewable biomaterials) (Sederoff et al. 2009; Groover 97 2007; Tsang et al. 2007). Thus, genomic basic research can provide both tool and resource 98 development, functional genomics tools in “omics” technologies that relate to high throughput 99 sequencing methods, bioinformatics, databases, and others, e.g., with respect to the development 100 of improved agricultural crop varieties and increased crop productivity using marker-assisted 101 breeding (see for example “Basic Research to Enable Agricultural Development, BREAD” 102 within the National Plant Genome Initiative, USA). Non-tree breeding related benefits from 103 genomics research include diagnostic tools to detect and map known and novel pathogens in real 104 time, provide risk assessments and support phytosanitation efforts. Such genomics-developed 105 tools can generate significant and immediate economic benefits ranging from a reduction in 106 losses due to forest diseases and protecting forest tree nurseries and natural forests from an 107 6 invasive pathogen to the support of Canadian forestry trade. Other opportunities for forest 108 genomics with respect to tree breeding and selection are by comparison considered long-term. 109 Due to the lack of a clear procedure to identify and justify research priorities, 110 “inconsistency of current funding mechanisms” for individual research projects that would best 111 support the forest sector strategy (Rank and Associates 2013) and a lack of economics analyses 112 supporting and driving research priorities, here, we attempted to (a) develop a robust typology of 113 forest sector genomics research initiatives relevant to Canadian application, (b) categorize each 114 initiative for its application potential (commercial and non-commercial, i.e. ecological/social, 115 including the estimated length of time distant from real application), and (c) demonstrate with 116 the three themes, silvicultural gain, insect resistance, and wood composition, the application of 117 modeling and economic analysis. Thus, the overall goal of the present study is to apply an 118 economics framework to genomics research to indicate which areas of genomics hold the most 119 promise for contributions to the forest sector. 120 Determining Thematic Areas in Forest Genomics 121 In order to support our assessment of the subject areas of funded forest genomics 122 projects undertaken thus far, we reviewed previous approaches taken to categorize the Canadian 123 projects, “tools for application” - markers, diagnostics, and enzymes (Terry Hatton, 124 unpublished), or “Healthy Forests” and “Productive Forests” (Rank and Associates 2013). Our 125 inventory of forest genomics projects comprised 22 Canadian 126 (http://www.genomecanada.ca/en/centres/) and 14 US projects (US Department of Energy 2013). 127 Although we are aware that there exist several ways to dissect project subject areas, ultimately 128 we felt that it was prudent to disaggregate forest genomics into different themes (or priority 129 7 areas). After a comprehensive review of all project focuses, we identified seven themes, as 130 summarized and defined in Table 1, that we feel best describe the project focuses to which all the 131 36 projects we had retrieved could be assigned and that are all relevant to the Canadian forest 132 sector (both Canadian and US based forest genomics projects: Table 2; Table S1): silvicultural 133 gain, adaptation, insect resistance, pathogen resistance, wood composition, wood ultrastructure, 134 and market access. We note here, however, that these existing themes within forest genomics 135 projects could not be identified for the ethical, environmental, economic, legal and social impacts 136 (GE3LS) project sections. GE3LS represent companion projects to most of the large-scale 137 genomics projects addressing elements of ethical, environmental, economic, legal and social 138 impacts of these genomics projects. We also observed here that, so far, no new emerging themes 139 could be identified, but we do suspect that technology transfer will become one of the key 140 themes in the very near future (see also below). 141 Table 2 represents the summary of the thematic areas used in other reports evaluating 142 forest genomics. We aimed to provide this comparison to show how well our seven themes align 143 with previous attempts to disaggregate genomics in a meaningful way and the extent to which 144 our approach can be regarded as the most comprehensive while at the same time avoiding 145 redundancy (Table 2). Another way of classifying these forest genomics projects would include 146 the four categories: (1) reduced losses (due to insects and pests; stress-resistance), (2) increased 147 growth, (3) wood quality, and (4) efficiency gains. Such a classification has its advantage when 148 general risk assessment is desired across broader focus areas within forest genomics projects as 149 almost every project focus aligns with almost every theme (Table 2 and Table S1). We included 150 “technology transfer” as a theme as one project has already implemented a technology transfer 151 committee. From this we assume that technology transfer is an emerging theme, which is aligned 152 8 with the community engagement priority areas outlined by the Genome BC strategy report 153 (Table 2). 154 Inventory of Forest Genomics Research Based on Thematic Areas 155 The thematic areas pursued by these projects are one way to organize the efforts in forest 156 genomics. These seven thematic areas (Table 2) constitute the basis of our forest genomics 157 research inventory. Table 3 provides the summary of all the project themes that characterize the 158 36 identified projects relevant to the Canadian forest sector and indicates the potential 159 application of the conducted research. The identified themes move along the supply chain from 160 the condition of the forest to forest products to forest products trade and markets. We have 161 chosen not to identify individual projects since our objective was not to target an individual 162 project but to present a comprehensive summary of all the genomics projects undertaken in 163 North American since 2002 with relevance to the Canadian forest sector. 164 Tables 2 and 3 also indicate the four basic tools for application of genomics as identified 165 in our thematic review: markers, diagnostics, viral proteins (here: insecticides), and enzymes. 166 The potential application of the tools is also summarized and it provides a further breakdown of 167 the themes (Table 3). We also indicated the number of projects undertaken within each theme for 168 both Canada and the USA; for Canada we also used available information to estimate what the 169 budget allocation per theme has been thus far (Table 3). The budget for all Canadian projects 170 adds up to a total of approximately $123 million (Table S2). In comparison to the Canadian 171 forest genomics initiatives, it is also interesting to note that all US initiatives focused on markers 172 and diagnostics as tools for application thus far along with an exclusive focus on “Productive 173 Forests” as the Genome category (Rank and Associates 2013); Table S1. 174 9 Table 3 further summarizes the purpose of application into three categories: upstream, 175 downstream and non-commercial (meaning a social and ecological focus in the forest genomics 176 research) applications. We used the symbols ‘x’ and ‘X’ to indicate the strength of the 177 application in any of those three categories. The next column is our estimate of the years to the 178 time when tools gained from the research could be applied. We chose to use three time periods: 179 (a) 1-5, (b) 6-10, and (c) 10+ years. The 1-5 years should indicate near-term opportunities of 180 application, i.e. these tools are close to commercial promise, in other words, they are largely 181 ‘private goods.’ Those thematic areas that are labeled 10+ years are those that represent mostly 182 fundamental research and the justification for such initiatives is based more on generating long-183 term ‘public good’ benefits; this is particularly evident for the subject area related to silvicultural 184 gain. Naturally, those subject areas labeled as 6-10 years are those closer to commercial 185 application and such shorter timelines of application were identified for certain initiatives related 186 to insect and pathogen resistance and were exclusively associated with risk assessment models. 187 Finally, looking at the combination of application purpose and the estimated time to 188 actual implementation in a genomics-informed operational forestry (“time to application”) does 189 indicate that very few commercial upstream initiatives will likely be applied in under 10 years 190 and since 2001, this is where the majority of the research funding has been allocated (Table 3). 191 Therefore, it might be more appropriate to consider them to be non-commercial in nature, since 192 the research is likely at the basic research stage and the ability to transform findings into 193 commercial application is simply unknown. For the commercial downstream thematic areas we 194 know that the funds allocated are relatively small and most initiatives undertaken so far are also 195 basic research. Nonetheless, since they are further down the supply chain, it does seem 196 reasonable to assume that these downstream applications, combined with advanced initiatives 197 10 related to chemistry and bio-pathways, could yield promising results. The non-commercial 198 application purpose has been to address longer term socio-ecological challenges such as climate 199 change adaptation. Nonetheless, these applications do have an economic impact, even if it is in 200 the very long term, but to our knowledge, no economic assessment results are publicly available. 201 Figure 1 provides a summary of the estimated timeline of application, and the categories 202 are again estimated to be short-term (1-5 years), medium-term (6-10 years) and long-term (10+ 203 years) for the various themes pursued in genomics research. Here, we also indicated the 204 percentage of the budget allocated for each specific theme within the graphic. Clearly, the 205 budgetary emphasis of research funding application has been on the longer term in the areas of 206 silvicultural gain and insect resistance; see also Table 3 and Table S2 for more details. In 207 general, it appears that the main focus of the research efforts was related to the condition of 208 forests defined as “Healthy Forests” and “Productive Forests” in Genome Canada terminology 209 (Rank and Associates 2013) and as such, the timeline for application in this category was, in 210 general, long-term (Figure 1). This can sometimes be justified given that the forests in Canada 211 are, at large, publicly owned and governments are tackling long-term issues such as timber 212 supply, biodiversity and community stability. The medium-term thematic areas are identified to 213 be non-commercial in nature, and with an emphasis on ecological and/or social issues. However, 214 it is noteworthy that initiatives which are short-term, i.e. their main focus is largely commercial 215 in nature and linked to forest products as opposed to forest condition, have received relatively 216 small budgets so far. 217 In concluding, our results obtained from the disaggregation of forest genomics and the 218 subsequent summary of individual thematic areas (Table 3) are in complete agreement with the 219 Forest Sector Challenges, Genomics Solutions report published by Genome Canada in December 220 11 2013 which generally states a “heavy focus on improved breeding opportunities” that represent 221 generational investments, while important, they can provide less short-term benefits (Rank and 222 Associates 2013). With this in mind, a re-thinking related to the strengthening of research 223 emphasis on short-term benefits for the forest sector has been recently advised (Rank and 224 Associates 2013). 225 Economic Analysis Review of Forest Genomics 226 To date, only few forest genomics initiatives have had an economic focus as part of their 227 research, consequently, the overall economic claims related to forest genomics thematic areas 228 were largely unsubstantiated. Here, we sought to explore themes where research in economics 229 did occur and found the most useful economic analysis in three themes: (a) insect resistance, (b) 230 silvicultural gain, and (c) wood composition. These themes cover commercial upstream and 231 commercial downstream analysis classes, respectively (Table 3). We also discuss the different 232 challenges that genomics is going to face in terms of its implementation and the economics of 233 genomics, particularly relevant to the commercial upstream applications (wood supply, tree 234 breeding). 235 Relationship of Genomics to Wood Supply 236 The economic studies on insect resistance and, associated with it, the timber supply were 237 led by Dr. Olaf Schwab, currently a senior economist with the Canadian Forest Service. Figure 2 238 indicates for the spruce weevil the potential impact of forest related genomics research on wood 239 supply (i.e. avoided losses). Given the long rotation ages of spruce trees, substantial benefits 240 from planting genetically improved weevil-resistant stock will only become relevant 250 years 241 from now (30% losses reduction). Moreover, the (Schwab et al. 2011) study clearly indicated 242 12 that the magnitude of the impact of planting genetically improved trees is driven largely by 243 assumptions on variables, such as discount rates, than by the loss of wood supply by the weevil. 244 Given the prevalence of neo-classical economics assumptions in forest decision making, discount 245 rates are the main factor that determines timber supply in the face of using weevil resistant stock. 246 For example, a 25% resistance increase and a 5% discount rate would translate into almost zero 247 dollars of benefit for avoided merchantable volume losses. This was identified as the most 248 unfavorable scenario. In contrast, a substantial 75% resistance increase and a 1% (i.e., lowest 249 assumed) discount rate could potentially translate into a $1.1 billion benefit due to the avoided 250 merchantable volume losses. Even planting large areas with weevil resistant trees only a small 251 portion of merchantable volume losses can be avoided, and, as mentioned, there is the long lag 252 time to consider. Thus, there are no immediate economic benefits to the forest sector unless we 253 include either the Allowable Cut Effect, or make the entries into landscapes where the avoided 254 losses were particularly important to maintaining timber supply through a pinch point in 255 available stands for harvest. Another uncertainty in the weevil-spruce model is the cumulative 256 merchantable volume losses (non-lethal damages) at the time of harvest. While, defect-free 257 lumber could still be produced or damaged lumber could be used otherwise as alternative 258 feedstock in bioenergy, the uncertainty makes investment planning difficult. Hence, there is a 259 need for a comprehensive cost-benefit analysis as large-scale investments into weevil-resistant 260 stock might bear a substantial risk (Schwab et al. 2011). 261 Cost Effectiveness of Genomics 262 The planting of genetically improved trees can have significant financial benefits, 263 provided that the costs associated with generating the planting material are not too high 264 (Petrinovic et al. 2009). Thus, to justify the application of genomics, the costs attached would 265 13 have to be assigned over a very large sale and further analysis of the benefits of the time savings 266 is warranted. Due to the long rotation cycles in Canadian forestry, conducting economics 267 analyses on “breeding-objective traits” is difficult (Ivković et al. 2010). A preliminary appraisal 268 of the costs and benefits with regards to genomics was recently conducted. These economic 269 studies connected to silvicultural gain were led by Dr. Nancy Gélinas, a professor of forest 270 economics at Laval University. She modified the methods in Dr. Harry Wu’s study on Australian 271 pine (Ivković et al. 2010) for Canadian boreal forest conditions, specifically with a focus on 272 volume and wood quality in Eastern white spruce. Accelerated breeding is able to cut the 273 duration of propagation and breeding phases by half (Resende et al. 2012b). Seedling production 274 costs are higher when genotyping and somatic embryogenesis (SE) are involved in seedling 275 production but will be largely offset by more volume, a shortened timeline and production of 276 higher value products. The integration of SE can optimize the benefits from the time gain for 277 rotation such that nine years can be saved to obtain improved seedlings compared to traditional 278 tree breeding (Gélinas unpublished). The optimal economic rotation age could also be reduced 279 by up to nine years for genetically improved white spruce (Petrinovic et al. 2009). Traditional 280 selection delays the rotation by 20 years. Genomic prediction, an approach originally developed 281 for dairy cattle breeding ((Meuwissen et al. 2001); (VanRaden 2008)), can substantially reduce 282 the entire tree breeding cycle, as the lengthy progeny testing phase is replaced by the prediction 283 of the total genetic value of the non-progeny tested population based on genotypic (DNA 284 marker) information (Grattapaglia and Resende 2011). Here, the effects of all available genome-285 wide genetic information are simultaneously tested on the trait of interest. This approach is 286 particularly valuable for low heritable traits and late expressing traits, as it allows for accurately 287 evaluating traits at an early age. Additionally, the application of genomic selections offers a new 288 14 dimension to selection through substantial increase in selection intensity and thus a 289 corresponding increase in the response to selection which is expected to be even greater than that 290 obtained through the application of genomic selection in traditional breeding programs (as poor 291 performers can be eliminated already at the zygotic embryonic stage). In summary, the 292 introduction of genomic selection with emphases on large selection intensity coupled with SE is 293 expected to produce gains far exceeding those obtained through traditional breeding (El-294 Kassaby, unpublished). Costs for whole-genome sequencing and genotyping have dropped 295 almost exponentially over the past 10 years (Poland and Rife 2012), yet phenotyping for 296 important wood characteristics is still tedious and costly (Porth et al. 2013b). While in traditional 297 tree breeding every single tested tree requires phenotyping, with genomic selection, in the end, 298 only a fraction of trees needs to be phenotyped for the proper development of accurate genomic-299 based prediction models as progeny testing can be omitted. Currently, there is no published bio-300 economic modeling for forest trees that compares the benefits of genomic selection to traditional 301 breeding. 302 Genomics to Products 303 The final examples of economic analysis relate to the potential of breeding for 304 lignocellulosic traits as well as improvement of value-added processing of wood fiber. The 305 economic studies for particular aspects of bioethanol production (commercial downstream) were 306 carried out by Dr. Catalin Ristea (Ristea 2014) and Dr. Jamie Stephen (Stephen 2013). 307 Using poplar as a feedstock, Ristea (2014) compared different land types for biomass 308 production (non-irrigated, irrigated; low, medium, high productivity) and demonstrated that for 309 the idle treatment scenarios the decrease in the net average area used was more substantial (Table 310 15 S3), highlighting the impact gained by improvements in both biomass yield and conversion 311 yield. 312 Stephen (2013) summarized the production costs for bioethanol that relate to 313 lignocellulosic feedstock and enzymes and further included opportunities for genomics research 314 to reduce such costs that relate to feedstock and enzymes (Table S4). For example, the reduction 315 in enzyme loading and of unproductive binding of the cellulase enzyme to lignin is considered 316 key in R&D for the optimization of cellulose hydrolysis. This can be achieved by development 317 of novel cellulase enzymes and custom designed enzyme cocktails for specific feedstocks as well 318 as an optimized feedstock with reduced lignin content which should facilitate feedstock pre-319 treatment and subsequently improve saccharification through reduced cell wall recalcitrance 320 (Van Acker et al. 2013; Rico et al. 2014; Wilkerson et al. 2014). In conclusion, these studies 321 (Ristea 2014; Stephen 2013) suggest that an integration of genomics has high potential in bio-322 processing improvement of wood fiber. 323 Conclusions 324 The present review on forest genomics attempted to improve rigor and structure to the 325 discussion over the use and application of genomic tools to the forest sector. Given that over 326 $123 million were spent since 2001, it is appropriate to reflect on what the focus has been in 327 terms of areas of research and its potential application. In order to lay the foundation for a 328 discussion on what the future areas of focus should be this review is intended to provide a 329 summary for decision makers. 330 So far, relatively little of the genomics budget has been allocated to economic analysis. 331 Given the preponderance of projects that could be considered basic research, it is appropriate, to 332 16 reflect on the degree to which future projects will have a different focus on commercial 333 application. 334 It is fair to say that thus far, the budgetary focus has been on the public forest resources, 335 not the private goods associated with private forest lands or forest products. However, at some 336 point, it needs to be assessed how the science of genomics can be associated more with potential 337 commercial benefits. Balancing resource allocation will become a great challenge as we move 338 forward with forest genomics research. It is important to highlight the new funding initiatives 339 created by Genome Canada and Genome BC that are rooted in involving the “end-user” as (an) 340 integral partner(s) in these funding opportunities. These initiatives have the potential to assess 341 the usefulness of and drive the immediate integration of genomics research to application. These 342 include the User Partnership Program (UPP), Genomic Applications Partnership Program 343 (GAPP) and the Strategic Opportunities Fund for Industry (SOFI) [www.genomebc.ca/research-344 programs/opportunities/current-funding-competitions/]. 345 Our overall goal was to develop an improved approach to the economic assessment of 346 forest genomics. We contributed to the discussion by: 347 1. Developing a typology of genomic projects. In our view, our typology is robust 348 and forward looking. 349 2. Categorizing the existing projects in terms of their potential application 350 (commercial up-stream; commercial down-stream; non-commercial, i.e. social and ecological) 351 and the length of time from real application. Clearly, one role of future GE3LS projects would be 352 to provide an assessment of the economic impacts for these potential applications. 353 3. Reviewing and commenting on the existing economic analysis conducted so far 354 under three themes: silvicultural gain, insect resistance and wood composition. 355 17 In many respects, the economic analysis will be challenging since time matters and 356 predicting the commercial application of basic research in economics is notoriously difficult for 357 all researchers concerned. Nonetheless it must be done. To make headway, it will be imperative 358 to develop a systematic approach and systems analysis to encourage a dialogue between 359 geneticists, economists and other scientists. 360 18 REFERENCES Aukema, B., Bohlmann, J., Bonfils, A.-C., Doucet, D., Elouafi, I., Erbilgin, N., Seguin, A., and Smith, S. 2009. Canadian Forest Health Genomics: Canadian Strengths Address Forestry Challenges. Canadian Genome Centres. Beaulieu, J., Doerksen, T., Clement, S., MacKay, J., and Bousquet, J. 2014. Accuracy of genomic selection models in a large population of open-pollinated families in white spruce. Heredity 113(4): 343-352. Bhalerao, R., Nilsson, O., and Sandberg, G. 2003. Out of the woods: forest biotechnology enters the genomic era. Curr Opin Biotechnol 14(2): 206-213. Costanza, A., and McCord, S. 2009. Forest Biotechnology and its Responsible Use [online] Institute of Forest Biotechnology. http://www.responsibleuse.org/resources/Forest_Biotechnology_and_its_Responsible_Use.html El-Kassaby, Y.A., and Lstiburek, M. 2009. Breeding without breeding. Genet Res 91(2): 111-120. Grattapaglia, D., and Resende, M.D.V. 2011. Genomic selection in forest tree breeding. Tree Genet Genomes 7(2): 241-255. Groover, A.T. 2007. Will genomics guide a greener forest biotech? Trends Plant Sci 12(6): 234-238. Holliday, J.A., Ralph, S.G., White, R., Bohlmann, J., and Aitken, S.N. 2008. Global monitoring of autumn gene expression within and among phenotypically divergent populations of Sitka spruce (Picea sitchensis). New Phytol 178(1): 103-122. Ivković, M., Wu, H., and Kumar, S. 2010. Bio-economic Modelling as a Method for Determining Economic Weights. Silvae Genet 59: 77-90. Kirst, M., Myburg, A.A., De Leon, J.P.G., Kirst, M.E., Scott, J., and Sederoff, R. 2004. Coordinated genetic regulation of growth and lignin revealed by quantitative trait locus analysis of cDNA microarray data in an interspecific backcross of eucalyptus. Plant Physiol 135(4): 2368-2378. Kumagai, K., Culver, K., and Castle, D. 2010. Public-Private Research Collaborations in Canadian Forestry Genomics: Knowledge Management and Innovation. The Integrated Assessment Journal 10(1): 15. McKown, A., Klápště, J., Guy, R., Geraldes, A., Porth, I., Hannemann, J., Friedmann, M., Muchero, W., Tuskan, G., Ehlting, J., Cronk, Q., El-Kassaby, Y., Mansfield, S., and Douglas, C. 2014. Genome-wide association implicates numerous genes underlying ecological trait variation in natural populations of Populus trichocarpa. New Phytol. doi: 10.1111/nph.12815. Meuwissen, T.H.E., Hayes, B.J., and Goddard, M.E. 2001. Prediction of total genetic value using genome-wide dense marker maps. Genetics 157(4): 1819-1829. Petrinovic, J.F., Gelinas, N., and Beaulieu, J. 2009. Benefits of using genetically improved white spruce in Quebec: The forest landowner's viewpoint. Forest Chron 85(4): 571-582. Poland, J.A., and Rife, T.W. 2012. Genotyping-by-Sequencing for Plant Breeding and Genetics. Plant Genome 5(3): 92-102. 19 Porth, I., and El-Kassaby, Y.A. 2014. Current status of the development of genetically modified (GM) forest trees world-wide: a comparison with the development of other GM plants in agriculture. CAB Reviews 9(8): 12. Porth, I., Klapste, J., Skyba, O., Friedmann, M.C., Hannemann, J., Ehlting, J., El-Kassaby, Y.A., Mansfield, S.D., and Douglas, C.J. 2013a. Network analysis reveals the relationship among wood properties, gene expression levels and genotypes of natural Populus trichocarpa accessions. New Phytol 200(3): 727-742. Porth, I., Klapste, J., Skyba, O., Lai, B.S.K., Geraldes, A., Muchero, W., Tuskan, G.A., Douglas, C.J., El-Kassaby, Y.A., and Mansfield, S.D. 2013b. Populus trichocarpa cell wall chemistry and ultrastructure trait variation, genetic control and genetic correlations. New Phytol 197(3): 777-790. Porth, I., White, R., Jaquish, B., Alfaro, R., Ritland, C., and Ritland, K. 2012. Genetical Genomics Identifies the Genetic Architecture for Growth and Weevil Resistance in Spruce. Plos One 7(9): e44397. Prunier, J., Pelgas, B., Gagnon, F., Desponts, M., Isabel, N., Beaulieu, J., and Bousquet, J. 2013. The genomic architecture and association genetics of adaptive characters using a candidate SNP approach in boreal black spruce. BMC Genomics 14(1): 368. Rank, D., and Associates. 2013. Forest Sector: Challenges, Genomic Solutions. Genome Canada. Resende, M.D.V., Resende, M.F.R., Jr., Sansaloni, C.P., Petroli, C.D., Missiaggia, A.A., Aguiar, A.M., Abad, J.M., Takahashi, E.K., Rosado, A.M., Faria, D.A., Pappas, G.J., Jr., Kilian, A., and Grattapaglia, D. 2012a. Genomic selection for growth and wood quality in Eucalyptus: capturing the missing heritability and accelerating breeding for complex traits in forest trees. New Phytol 194(1):116-128. Resende, M.F.R., Jr., Munoz, P., Acosta, J.J., Peter, G.F., Davis, J.M., Grattapaglia, D., Resende, M.D.V., and Kirst, M. 2012b. Accelerating the domestication of trees using genomic selection: accuracy of prediction models across ages and environments. New Phytol 193(3): 617-624. Rico, A., Rencoret, J., del Rio, J.C., Martinez, A.T., and Gutierrez, A. 2014. Pretreatment with laccase and a phenolic mediator degrades lignin and enhances saccharification of Eucalyptus feedstock. Biotechnol Biofuels 7(1): 6. Ristea, C. 2014. Modeling the Net Greenhouse Gas Balance of Projects that Display Gasoline with Wood Ethanol from Short Rotation Tree Plantations. Unpublished draft of PhD dissertation. Forestry, UBC, Vancouver. Schwab, O., Maness, T., Bull, G., Welham, C., Seely, B., and Blanco, J. 2011. Modeling the timber supply impact of introducing weevil-resistant spruce in British Columbia with cellular automata. Forest Policy Econ 13: 61-68. Sederoff, R., Myburg, A., and Kirst, M. 2009. Genomics, domestication, and evolution of forest trees. Cold Spring Harbor symposia on quantitative biology 74: 303-317. Stephen, J. 2013. The Variability of Lignocellulosic Ethanol Production as a Business Endeavour in Canada. Unpublished PhD Dissertation. Forestry, UBC, Vancouver. Street, N.R., Skogstrom, O., Sjodin, A., Tucker, J., Rodriguez-Acosta, M., Nilsson, P., Jansson, S., and Taylor, G. 2006. The genetics and genomics of the drought response in Populus. Plant J 48(3): 321-341. Tsang, A., Douglas, C., Mohn, B., Bocking, S., Einsiedel, E., Facchini, P., Larouche, A., Mansfield, S., Page, J., Regan, S., Singh, J., Smith, M., and Weselake, R. 2007. 20 Genome Canada-Strategic Research Theme: Securing Canada's Future Bio-Based Economy through Genomics. The Biofuels and Bioproducts Genomics Working Group. Biofuels and Bioproducts Genomics. US Department of Energy, O.o.S. 2013. Plant Feedstock Genomics for Bioenergy Joint Awards 2006-2013 [online]. Edited by National Institute of Food and Agriculture United States Department of Agriculture, http://GenomicScience.energy.gov/researh/DOEUSDA/. Van Acker, R., Vanholme, R., Storme, V., Mortimer, J.C., Dupree, P., and Boerjan, W. 2013. Lignin biosynthesis perturbations affect secondary cell wall composition and saccharification yield in Arabidopsis thaliana. Biotechnol Biofuels 6(1): 46. VanRaden, P.M. 2008. Efficient Methods to Compute Genomic Predictions. J Dairy Sci 91(11): 4414-4423. Verta, J.-P., Landry, C.R., and MacKay, J.J. 2013. Are long-lived trees poised for evolutionary change? Single locus effects in the evolution of gene expression networks in spruce. Molecular Ecology 22(9): 2369-2379. Wilkerson, C.G., Mansfield, S.D., Lu, F., Withers, S., Park, J.Y., Karlen, S.D., Gonzales-Vigil, E., Padmakshan, D., Unda, F., Rencoret, J., and Ralph, J. 2014. Monolignol Ferulate Transferase Introduces Chemically Labile Linkages into the Lignin Backbone. Science 344(6179): 90-93. 21 TABLES Table 1: Thematic themes identified among 36 forest genomics projects Theme Description Market access Research allowing (immediate) diagnosis of tree health (pathogen diagnostics). Climate change adaptation Research on assessment of the adaptive potential including stress-resistance in forest tree populations; knowledge-based directives (assisted migration). Insect resistance Research on bark beetles, weevils, budworm; development of insecticides. Pathogen resistance Research on plant pathogens, especially fungi and insect associated pathogens (e.g. MPB associated blue stain fungus). Wood composition Research focused on the percentage cellulose versus lignin, but also on altered lignin structure where applicable (easier breakdown of lignocellulosic feedstock for bioethanol production). Wood ultrastructure Research on fiber properties important for pulp and paper as well as timber industry; wood density is also related to wood ultrastructure and an important determinant for yield (e.g. biofuel lignocellulosic feedstock). Silvicultural gain (breeding and selection gain) Research on tools such as MAS (marker assisted selection); research aims at identifying primarily genetic markers underlying traits of interest; development of diagnostic tools for breeding targets to facilitate market access. It also refers to general efforts to shorten breeding cycles and increasing selection intensity in trees (genomic selection, e.g.); conditions to support improved gain (soil microbial conditions; hybrid breeding); access to improved management tools. 22 Table 2: Large scale and follow-up Canadian research projects in forest genomics Theme Genomic Tools for Application Genome Canada Category Genome BC Category Projects Focus Silvicultural gain (breeding and selection gain) markers; diagnostics Healthy Forests, Productive Forests Healthy Forests, Productive Forests Wood Quality; Reduce Losses; Increased Growth; Efficiency Gain Adaptation markers; diagnostics Healthy Forests, Productive Forests Healthy Forests, Productive Forests Wood Quality; Reduce Losses; Increased Growth; Efficiency Gain Increase insect resistance markers; diagnostics Healthy Forests Healthy Forests Wood Quality; Reduce Losses; Increased Growth; Efficiency Gain Increase pathogen resistance markers; diagnostics Healthy Forests, Productive Forests Healthy Forests, Productive Forests Wood Quality; Reduce Losses; Increased Growth; Efficiency Gain Wood composition markers; enzymes Healthy Forests, Productive Forests Downstream Application Wood Quality; Reduce Losses; Increased Growth; Efficiency Gain Wood ultrastructure markers; diagnostics Productive Forests Downstream Application Wood Quality; Reduce Losses; Increased Growth; Efficiency Gain Market access diagnostics Healthy Forests Healthy Forests Reduce Losses (Technology Transfer) - - Engaging Communities - 23 Table 3: A comparative summary of classification of forest genomics projects Theme Tools for Application Potential Application of the Research (can be more than one row per theme) Total number of Canadian Projects per theme Approved budget of Canadian Projects (million) Number of US projects per theme Application Purpose Estimated time to application (years), 1-5, 6-10, 10+ Commercial up-stream Commercial down-stream Non-commercial (social and ecological) Silvicultural Gain Markers Identify fast growing varieties for biofuel feedstock 7 $2.5 8 x 10+ Develop genomic methods and tools to enhance genetic selection and breeding $22.7 x 10+ Develop environmental genomic tools for improving forest management practices $3.9 x 10+ Identify desired mutants based on genotype of targeted genes to generate novel varieties suitable as bioethanol x 10+ Develop specialized bioenergy tree cultivars x 10+ Develop varieties with superior biomass feedstock potential x 10+ Understand how endophytic bacteria alter plant growth and productivity, to ultimately manipulate plant performance for feedstock production x 10+ Improve hybrid breeding of tree varieties that are used to produce wood and bioenergy x 10+ Improve yield by capturing hybrid vigor x 10+ Understand phytochrome-mediated responses to competition to maximize carbon capture per unit of land area for increased biomass production x 10+ Diagnostics Develop genomic tools to enhance forest productivity $3.8 x 10+ Identify germplasm with unique genotypes and increased biomass yields for tree breeders X 10+ Apply hyper-accelerate breeding using genome-wide selection, develop new breeding strategies X x 1-5yrs Adaptation Markers Assess microbial diversity for 7 $5.2 4 x 10+ 24 Theme Tools for Application Potential Application of the Research (can be more than one row per theme) Total number of Canadian Projects per theme Approved budget of Canadian Projects (million) Number of US projects per theme Application Purpose Estimated time to application (years), 1-5, 6-10, 10+ Commercial up-stream Commercial down-stream Non-commercial (social and ecological) sustainable use of forest biomass resource Investigate tree interactions with soil microbiome x 10+ Investigate plant performance in association with microbial communities x 10+ Diagnostics Investigate physiological traits for adaptation and biomass productivity $2.2 x 10+ Assess the adaptive potential in forest tree populations; provide knowledge-based directives (on assisted migration) $5.7 x X 1-5yrs Generate genomic tools to study mycorrhizal ecology x 10+ Develop varieties that are abiotic stress tolerant and grow on marginal land x 10+ Develop robust biomass productivity under marginal conditions x 10+ Provide knowledge about dormancy induction x 10+ Insect Resistance Markers Develop and integrate genomics for ecological risk models 7 $7.8 0 X 6-10yrs Identify forest health markers to support breeding programs $25.3 x 10+ Viral proteins Generate insecticides using naturally occurring insect viruses $4.8 X 1-5yrs Diagnostics Increase insect resistance $3.8 X 10+ Pathogen Resistance Diagnostics Identify and monitor forest pathogens using DNA-based diagnostic tests 5 $2.1 0 X X 1-5yrs Markers Develop and integrate genomics for ecological risk models $7.8 X 6-10yrs Develop strategies to more quickly detect and monitor rust, and to more effectively prevent infection X 10+ Determine pathogen species, identify pest susceptible genotypes in breeding trials $2.5 X 10+ 25 Theme Tools for Application Potential Application of the Research (can be more than one row per theme) Total number of Canadian Projects per theme Approved budget of Canadian Projects (million) Number of US projects per theme Application Purpose Estimated time to application (years), 1-5, 6-10, 10+ Commercial up-stream Commercial down-stream Non-commercial (social and ecological) Wood Composition Markers Develop allelic markers in marker assisted breeding for accelerated feedstock improvement 4 $4.7 6 x 10+ Generate knowledge on genomics of wood formation $5.4 x 10+ Identify alleles with breeding values x 10+ Generate knowledge regarding the potential of the protein-protein interactions relevant to biomass production x 10+ Generate knowledgebase about genes for effective manipulation of lignocellulosic traits to facilitate ethanol production x 10+ Identify biochemical functions of acyltransferases in polysaccharide acetylation, lignol biosynthesis, and phenolic compound modification x 10+ Diagnostics Generate cultivars to produce high energy yields ready for deployment X x 1-5yrs Enzymes Identify fungal enzymes for breakdown of wood $2.2 x x 1-5yrs Isolate and identify novel enzymes for biorefining to cellulosic ethanol 4 0 x x 1-5yrs Wood Ultrastructure Markers Develop allelic markers in marker assisted breeding for accelerated feedstock improvement $4.7 x 10+ Diagnostics Develop diagnostic markers associated to high stiffness and dimensional stability $3.8 x 10+ Market access Diagnostics Identify and monitor forest pathogens using DNA-based diagnostic (phytosanitary status) tests 1 $2.1 0 X X 1-5yrs 26 FIGURE CAPTION Figure 1: Graphic summary of the emphasis of the funded genomic projects in Canada based on the estimated time to application of Canadian projects per defined theme. Figure 2: Effect of genetic resistance to the spruce weevil on merchantable losses (Schwab et al. 2011) 27 Supplementary Material Table S1. Inventory of thematic areas: 14 US research projects in Forest Genomics Table S1 represents the number of USA projects that fall under the 7 themes defined in this study. Also, these projects were aligned with four categories: (1) reduced losses (due to insects and pests; stress-resistance), (2) increased growth, (3) wood quality, and (4) efficiency gains as the project focus and with “Productive Forests” as the Genome category. Theme Tools for Application Category Project Focus Silvicultural gain (breeding and selection gain) markers; diagnostics Productive Forests Efficiency Gain, Wood Quality, Increased Growth, Reduce Losses Adaptation markers; diagnostics Productive Forests Efficiency Gain, Reduce Losses Increase insect resistance - - - Increases pathogen resistance - - - Wood composition markers; diagnostics Productive Forests Efficiency Gain, Wood Quality, Increased Growth Wood ultrastructure - - - Market access - - - 28 Table S2. Estimated time to application and percentage of allocated budget of Canadian projects Investment into 22 Canadian projects was merged together by defined 7 themes in Table S2. The time of application categories are estimated to be: short-term: 1-5 yrs, medium: 6-10 yrs and long-term: 10+. The amount of budget was equally divided into number of areas it covered, if a project covered more than one theme. See also Figure 1 for a graphical representation of this table. Theme Tools for Application Estimated time to application (years), 1-5, 6-10, 10+ Approved budget of Canadian Projects (million) Budget (%) Silviculture Gain Markers 10+ $2.5 2.03 10+ $22.7 18.46 10+ $3.9 3.17 Diagnostics 10+ $3.8 3.09 Adaptation Diagnostics 10+ $5.2 4.23 Markers 10+ $2.2 1.79 1-5yrs $5.7 4.63 Insect Resistance Markers 6-10yrs $7.8 6.34 10+ $25.3 20.57 Viral proteins 1-5yrs $4.8 3.90 Diagnostics 10+ $3.8 3.09 Pathogen Resistance Diagnostics 1-5yrs $2.1 1.71 Markers 6-10yrs $7.8 6.34 10+ $2.5 2.03 Wood Composition Markers 10+ $4.7 3.82 10+ $5.4 4.39 Enzymes 1-5yrs $2.2 1.79 Wood Ultrastructure Markers 10+ $4.7 3.82 Diagnostics 10+ $3.8 3.09 Market access Diagnostics 1-5yrs $2.1 1.71 SUM $123 100 29 Table S3: C-BOS model results by land unit scenario: improvements in both biomass yield and conversion yield; idle treatment allowed (Ristea 2014) Three of the land types were non-irrigated (LU1, LU3, and LU5) and three were irrigated (LU2, LU4, and LU6). For the non-irrigated types, LU1 had high productivity, LU3 medium productivity, and LU5 low productivity. Similarly for the irrigated types, LU2 had a high productivity, LU4 medium productivity, and LU6 low productivity. Land unit type Max annual area Net average area Site prep. unit cost Land rent unit cost Biomass production unit cost Biomass transport unit cost Harvested biomass (stem+bark +branch) [ha/yr] [ha/yr] [$/t] [$/t] [$/t] [$/t] [t] LU1 0 0 0 0 0 0 0 LU2 325 144 0 11.58 93.21 45.97 166,670 LU3 71,443 59,288 0.29 12.84 77.11 42.89 49,647,185 LU4 0 0 0 0 0 0 0 LU5 14,612 4,515 1.00 12.47 94.41 28.67 2,919,349 LU6 0 0 0 0 0 0 0 30 Table S4: Lignocellulosic ethanol production cost for base case 200,000 bdt/yr biomass processing facility (modified after Jamie Stephen (Stephen 2013)) FPU…filter paper units; FPU g-1 protein: protein activity bdt…bone dry ton † total costs for enzymes: at 600 FPU g-1 protein activity, 20 FPU g-1 cellulose, assuming 42% feedstock cellulose content, and yield of 250 L EtOH bdt-1 biomass requires 56 g protein per L EtOH; 10 FPU g-1 cellulose loading results in $4.40 kg-1 protein Facility Overview Research Opportunities for Cost Contribution Reduction (potential Genomics contribution) Plant Capacity 75.1 ML yr-1 - Ethanol Yield 321 L EtOH bdt-1 Wood quality (composition and ultrastructure) Unit Installed Cost $2.15 L-1 at yearly capacity - Total Capital Cost $161,500,000 - Feedstock Feedstock Cost $95.87 bdt-1 Improved productivity of woody feedstocks through selective breeding or genetic engineering Non-Feedstock Operating Costs Enzymes $2.20 kg-1 protein†; 600 FPU g-1 protein; 20 FPU g-1 cellulose Identification or development of novel cellulase enzymes; custom designed enzyme cocktails for specific feedstocks Revenues Specific revenue (year 1) $0.59 L-1 Co-product credit (lignin and C5 sugar-derived products)