[Note: these are rough notes taken down during the workshop itself by Anya Hinkle with contributions by Russ Chapman. These notes are intended for archival purposes; please excuse the rough edges -- we would appreciate getting feedback from anyone about errors or omissions in this document. They have been circulated to all participants for comment, but please don't quote people directly]

DEEP GREEN and DEEP GENE Workshop
17 FEB 2001
Westerbeke Ranch Conference Center, Sonoma, California

See also the meeting summary

9:25 am
1) Brent Mishler made opening remarks followed by introductions of all 36 participants.

9:39 am
2) Brent Mishler:
a) commented on the schedule of events and agenda for the day
b) reported on Friday press briefing at AAAS in San Francisco (more than 100 reporters present) and symposium (standing room only)
c) commented on the history of Deep Green and original grant in 1994 to coordinate work among labs in phylogenetics. Deep Green has been mentioned favorably by NSF staff including the Director Rita Colwell as an exemplary collaboration, and even featured in NSF budget requests to Congress to highlight the coordination of research among collaborators from different fields/institutions. Three related, new initiatives have come from Deep Green so far, leading to lots of coordination funds that can be used in part to develop additional grant proposals for lab research funds.
1) Deep Gene - bringing together phylogenetics and genomics - a Bay Area Biosystematists meeting held at Standford in May 2000 led to an NSF RCN proposal that was funded; need to decide just what will be done with the funds
2) Deep Time - a second NSF RCN headed by Doug Soltis to integrate the fossil record into molecular phylogenetics. - (the PDF file of both proposals were emailed to participants)
3) Biocomplexity - An NSF incubation grant for $100K, to begin collaborations on genome evolution at the chromosome level; Pam Soltis PI
d)Future goals: It would be fine to have additional "spin-off" proposals coming from this current activity with the idea of using the ideas generated in funded meetings (like today's) to stimulate proposals to fund actual bench work
e) Future goals: We need and want to incorporate more people in this activity, not just the "core participants" listed in the proposal.

9:48
3) Roundtable on Current Research Activities in Participants' Laboratories and Their Directions for the Future:

a) Rick McCourt: His research focuses on Charophycean algae in cooperation with Chuck Delwiche and with paleontologists in France. He is interested in comparative genomics of Zygnematales using chloroplast and mitochondrial genomes. Major groups of interest to him are Charophytes, a low diversity group but which have a good fossil record, and conjugating green algae (Zygnematales) with about 5000 species. He is interested in the evolution of algae and their relationship to the land plants. Early efforts focused on phylogenetics and rooting of land plant phylogeny, and now he is getting more into comparative genomics, and the use of genes present in land plants, such as MADS box genes and others, to see what they are doing and following them up through the phylogeny.

b) Charles Delwiche: Rick McCourt described their collaborative work on systematics and taxonomy of basal streptophytes. In addition to this, the Delwiche lab also works on the evolutionary origin of chloroplasts, and for this project they are currently performing an EST study of photosynthetic dinoflagellates. One specific problem concerning streptophyte phylogeny Delwiche is interested is the relationship of Mesostigma to streptophytes, and in particular its relationship to Chaetosphaeridium. Chloroplast genes put Mesostigma at base of the streptophyte clade but if you look at rRNA Mesostigma pairs with Chaetopshaeridium. In most particulars the chloroplast phylogeny is consistent with morphological analyses, but the rRNA data show something quite different. Close analysis reveals that there are two conflicting signals in the ribosomal data. While the ribosomal data do include some signal that is consistent with the chloroplast data, there is a second, conflicting signal that links Chaetosphaeridium with Mesostigma. This may be the result of convergent evolution of the ribosomal sequences, perhaps because of independent loss or modification of one of the ribosomal proteins. Delwiche is looking at the recently determined crystallographic structure of a bacterial ribosome to test this hypothesis, and suspects that Chaetosphaeridium and Mesostigma may have independently lost the ribosomal protein s8.

c) Chris Somerville: His interests have been in chemical diversity in plants in for many years (this program is more or less coming to an end). He comments on some of the results on lipid biochemistry obtained over the years such as how fatty acid diversity arose. He found the enzymes that modify fatty acids, by changing as few as 4 amino acids they were able to change major features (i.e. double bond or hydroxyl). He sought an understanding of how one chemical center, through small geometric changes, can generate many different lipids. Another result relates to how plant cell walls are made using Arabidopsis beta galactyltransferases with some similarity to cellulase synthetase. They make mutations in most of the genes, characterize the phenotype, and try to learn how they've evolved to make a wide variety of polymers. They think that a large family of enzymes has evolved using similar chemistry to polymerize sugars. Maybe a thousand genes are involved in making cell walls but only three are currently known; thus they are working on key enzymes via reverse genetics to identify key enzymes and explain how they work. Other work in the lab focuses on embryogenesis and looking at mutants with altered embryogenesis using the genome sequence of Arabidopsis. Today they can get from enzyme to gene (in "good" cases) in six weeks (six postdoc weeks as opposed to up to 6 postdoc years a few years ago). For Arabidopsis there is a rough idea of gene identification for about half the genome by sequence similarity but we need to understand what the other half of the genes (currently with unknown function) actually do. They have fairly significant funding to find out what the gene functions are. There are some 200,000 secondary metabolites in the angiosperms yet there are not enough genes to explain all of the chemical diversity that exists. Rapid evolution of function must require some kinds of rules for understanding how different classes of metabolites evolve.

d) Eric Roalson: (Liz Zimmer's lab) The Zimmer lab is interested in molecular phylogenetics and evolution and is currently working in the Gesneriaceae, Asteraceae and Onagraceae (the latter with funding from a new Mellon Grant). We are also working in collaboration with Judith Skog on fern phylogenetics. Eric Roalson himself is working on the integration of developmental genetics with phylogeny, and studying spur development in flowers in the Gesneriaceae (African violet family).

e) Dave Erickson: (Liz Zimmer's lab) Dave is looking at outcrossing and epistatically interacting genes (the number of genes involved and their relative effects). Using QTL mapping and biometrics, he expects an outcrosser to have the advantageous genes linked in contrast to an inbreeder, where the advantageous genes may be distributed throughout the genome. He is estimating the number of genes and how mating structure affects gene order on chromosomes.

f) Michael Zanis: (Soltis' lab) The lab is continuing to work on angiosperm phylogeny looking at basal angiosperms (e.g. Amborella and water lilies). They report continued strong bootstrap support for Amborella as sister to all other angiosperms, but cannot consistently reject all alternative hypotheses with the 11-gene data set they have been using. They have been adding full length 26S rDNA and matK sequences to the angiosperm dataset. They are also involved in the Floral Genome Project with several other investigators. They are interested in floral development (MADS box genes in basal angiosperm lineages and B class genes in Nuphar and Illicium). Also they are looking at population genetics and evolution of single copy and low copy genes in rare and common species. He reports on two grants they have going currently. One supports Biocomplexity work on the role of genome doubling in diversification of plants and genome evolution following genome duplication. Empirical work is focused on chromosomal reorganization in recent polyploids. The (other) Deep Time grant integrates fossil and extant plant data sets.

g) Shauna Somerville: Research in Shauna Somerville's lab focuses on questions in plant-pathogen interactions such as "Why are some plants resistant?" and conversely "How do susceptible plants contribute to disease development?" Most of the recent work in the Somerville lab is directed towards understanding the host components that contribute to susceptibility, using a mutant approach. In addition to the projects on host susceptibility, the Somerville lab is also studying host responses to foreign or nonhost pathogens using Arabidopsis thaliana as a host and the barley powdery mildew, as the nonhost pathogen. The final project is the Somerville lab relates to the use of DNA microarrays to study global patterns of gene expression in Arabidopsis. This technology, when coupled to gene expression databases, can be used to develop "expression signatures" characteristic of plant responses to attack by various pathogens for example.
Web page: http://carnegiedpb.stanford.edu/shauna/shauna.html

h) Steve Tanksely: His work focuses on plant variation. His short term evolution project is how plants developed large fruits, e.g. in Solanaceae (edible tomatoes from very small ancestor plant). His research asks questions about how many genes are involved in the process and what to they do? He used a cross between species and then used QTL mapping to locate the genes for fruit size and shape in tomato and found 25 loci that seem to account for the variation in tomato. They have isolated a gene involved in a negative regulatory function. He is also interested in what kinds of natural mutations occur (e.g. many small changes in the regulatory region). One example is a change in stigma insertion to accommodate selfing which is a single locus effect. Another is the gene for fruit size is also a regulator of cell division to which mutation reduces amount of gene product (a negative regulator of cell division). It also shows a very distant similarity to a RAS protein, a cancer - related gene in humans. What drives morphological evolution? Eggplant has been studied and a similar gene involved in fruit enlargement coincides with the tomato. His long term evolutionary question is how does genome structure evolve over long periods of time (i.e., compare different families separated in time)? Within the genomes of the two groups, he would like to use QTL mapping to locate the genes for fruit size and shape in tomato and eggplant. He needs to find highly conserved single copy genes to compare among plants. The goal is to understand arrangement and synteny among genomes.

i) Daphne Preuss: Her lab is interested in pollination pathways involving proteins located on the surface of pollen. In order to characterize the surface proteins they use mutants. Two groups of pollen coat proteins have been purified and sequenced: glycine-rich oleosins and extracellular lipases. One question is how do the proteins vary from species to species? Does the diversity of these proteins result in the ability to mate with stigmas of an appropriate genotype? They would like to extend their work beyond Arabidopsis. Another project is the characterization of centromeres, a region that is difficult to sequence. Five centromeres have been sequenced in Arabidopsis - we now have more information on plant centromeres than for human and other animal centromeres! One question is which sequences within the centromere are most important to centromere functionality? These should be the most conservative regions in the centromere sequences. Some of the repetitive sequence regions may be involved in condensation of the DNA. There seems to be signature sequences in these genes for the species and for the ecotypes. Sarah Hall is working on genes from different ecotypes and looking at polymorphisms. Conservative sites may be protein binding sites. An example is a 180 bp repeat, similar to the human alphasatellite178 bp repeat(180 is the length needed to wrap a nucleosome). Since each organism has a signature sequence, they are interested in finding out if the signature sequence plays a role in function.

j) Paul Wolf: His interests are in fern phylogeny and more recently extending to relationships of vascular plants (in conjunction with Kathleen Pryer). He is currently interested in resolving deep land plant branches using genome sequence and structural data. He is looking at how genomes evolve and how rare and/or complex rearrangements can be used as phylogenetic markers. The initial focus is with chloroplast genomes. Questions to be addressed include the resolution of ambiguity and controversy regarding which extant group is the most basal lineage of land plants, which groups are sister to the land plants, and the positions of critical taxa with a long fossil record such as the Equisetophyta. His current goal is the Biocomplexity proposal
to isolate and sequence about twenty chloroplast genomes. He is working on developing techniques for analysis and visualization that can be used to look at the genome on all scales from large scale inversions to small indels and single base substitutions. These can be used both for visualization and characterization of evolutionary changes. Dick Olmstead has detected many phylogenetically-informative INDELS in plastid sequences, that were useful at different levels. Ray Cranfill has also found length differences that characterize the moniliforms (ferns and some of their "allies"). Finally, he is interested in comparing blue-green "algal" genome sequences to chloroplast genomes to examine major rearrangements and trends in gene movement across genomes since the original endosymbiosis.

k) Yin-Long Qiu: His research follows two lines of inquiry. He began by sequencing a single gene (rbcL) to look at the phylogeny of basal angiosperms with extensive taxon sampling but without robust resolution. He then went on to look at the genomic structure of organellar genomes. He focused on three introns in the mitochondrial genes cox2 and nad1, and found that the gain of these introns as evidence for the basal position of liverworts vis-a-vis hornworts in land plants. Characterization of more mitochondria introns in his lab and other labs supports the basal placement of liverworts. Currently, they are sequencing eight genes across 150 land plants as well as the mitochondrial genomes of a hornwort and a moss, to resolve several contentious issues in land plants phylogeny. They have also sequenced five genes from mitochondrial, chloroplast, and nuclear genomes and identified Amborella, Nymphaeales, Illiciales-Trimeniaceae-Austrobaileya (ANITA) as the earliest extant angiosperm lineages.

l) Michael Sanderson: He and Marty Wojciechowski are currently working on a difficult genus of angiosperm, Astragalus. His interest is in using genomics and phylogenetics to understand relationships in crown groups with recent divergence times. This requires loci and introns from the nuclear genome since such events would be too recent for the chloroplast genome. He is looking for nuclear spacers, introns and third position regions for variability that can be used as characters for a phylogeny. Because of the small amount of sequence divergence (2-3%) it will require extensive sequencing of many genes and introns.

m) Yangrae Cho (postdoc in Virginia Walbot's lab): The Maize Gene Discovery Project has been funded to this lab for 2 years to support 1) 50,000 EST sequencing, 2)Transposon Tagging and Mutant Archiving, 3) cDNA microarray fabrication. Currently, over 76,000 ESTs have been sequenced and deposited to GenBank and EST clones have been distributed to diverse institutes, including individual researchers upon requests. Transposon tagging, mutant archiving and microarray fabrications are ongoing efforts. Three different issue specific microarrays were fabricated at Arizona University and they have been distributed to maize community. In the near future, 5 additional tissue specific microarrays and unigene (over 20,000 probes) microarrays will be fabricated. All three categories of work will be continued for coming three years. In addition to this initial effort, we are exploring diverse applications of genomic tools to evaluate genome annotation and to investigate exon splicing. Oligo microarray method is one of many approaches. Our interest in RCN is in two major aspects; evolutionary mechanisms of genome diversity and biochemical pathways.

n) Claude DePamphilis: His interests are in parasitic plants, evolutionary genomics, and the use of mitochondrial sequences in evolution and phylogenetics. He has completed a 2-gene mitochondrial phylogeny (note that parasites have very rapid rates of gene evolution for plastid and nuclear genes, and often very reduced plastid genomes) of over 150 seed plants and angiosperms. Mitochondrial Cox-1 and atpA together recover a phylogeny very similar to the recent plastid and nuclear phylogenies, including recovery of the 40 APG orders. In addition, there was clear resolution for many parasitic groups that were previously unplaceable. He has been using the RASA approach to try to improve phylogenetic analysis. In mitochondrial phylogeny studies it is important to remember that RNA-editing is important; mitochondrial genes seem to be subject to retroprocessing which has interesting implications. The other area in which they work is the evolutionary origin of flowers and the floral developmental program. Several labs (eg., Solti, Ma, Theissen, Albert, Frohlich, Tanksley) were interested in similar questions and recently proposed a joint plant genome project. This will primarily involve dense EST sequencing and high throughput expression analysis of genes expressed during early flower development in diverse angiosperms, and phylogenetic analysis. The EST approach will generate many nuclear genes for plant phylogenetic and evolutionary studies. We have linked this project with another recent proposal led by Brent Mishler to systematically identify homologs of genes involved in reproductive processes in ferns and bryophytes.

o) Kathleen Pryer: Kathleen has had a long-term interest in the systematics and phylogenetics of ferns, horsetails, and other basal tracheophytes (e.g., Osmunda, Equisetum, Psilotum, Lycopodium). In a recent publication that appeared in Nature, Pryer and her colleagues assembled and analyzed five combined data sets -- from morphology and from four genes -- showing that horsetails together with ferns form a major group of vascular plants, that they are the closest living relatives to seed plants, and that they have had a long evolutionary history separate from seed plants. This refutes a general idea
that many botanists have, which is that ferns and horsetails are intermediate transitional stages in plant evolution that eventually led to seed plants. This deep split in vascular plant evolution has only been proposed once before by Paul Kenrick and Peter Crane, on the evidence of a single anatomical character. Efforts to promote developmental and genomic research on model systems in the horsetail-fern clade (for example, Ceratopteris) will likely lead to an improved understanding of vascular plant development and evolution. Pryer is beginning to look at
mitochondrial and other nuclear genes to further resolve higher-level relationships among non-seed tracheophytes. She is also collaborating with paleontologists to integrate morphological data from fossil ferns into phylogenies based on morphological, developmental, and molecular data. Her work on aquatic heterosporous ferns is focused on better understanding the evolution of heterosporic phenomena in vascular plants. The heterosporous genus Pilularia has an exceptionally low chromosome number for ferns and an astonishingly fast life cycle, indicating that it might also be an excellent taxon to target for future model organism/genome sequencing
studies. In addition, these heterosporous ferns have indurate, protective sporocarps that may be useful in physiological studies directed at elucidating the mechanisms of dessication tolerance in land plant evolution studies.

p) Toby Kellogg: The big question in which they are interested in is morphological diversification - What changes in the genes are responsible for such diversity of form? Model systems have genome maps and/or sequences, mutants, stocks, forward and reverse genetics, transformation, multiple ways to assess gene expression (antibodies, in situ, reporter constructs). Such tools are powerful for model systems, and are still available as long as you can do crossing. Once you move outside the sterility barrier, you have lost most of your tools (i.e. mutants, stocks, forward/reverse genetics, transformation). The grass family provides additional tools for understanding morphological diversification as well as providing some models (rice has recently been sequenced) because there is good alpha taxonomy, a good phylogeny, genome maps, and information on cytology and physiology. Can we correlate molecular evolution in expression of particular genes with changes in morphology? To correlate gene expression with morphological change we need to have a better characterization of the phenotype (developmental morphology and histology). This includes the identification of candidate genes based on the following criteria:
1) genes that are developmentally important - are they also evolutionarily important?
a) knotted1 (Tony Verboom)
b) leafy-Andrew Doust
c) B&C class genes
2) genes with mutant phenotypes which naturally occur in other species (tassel seed2 - Simon Malcomber)
3) genes whose expression pattern is known to vary among the taxa of interest (such as photosynthetic genes -Sinha & Kellogg 1996)
4) QTL studies on non-models with interesting phenotypes (S. italic X S. viridis - Deros & Gale and Doust & Kellogg).
Future challenges involve developing more efficient methods for screening for gene expression. She addresses a need for rapid primer design (aided by genome sequences) and universal transformation systems.

q) John Doebley: His research has two foci: 1) The first concerns the evolution of plant form in maize and its wild relative teosinte. Here, the Doebley Lab is seeking to identify, clone and characterize the genes that were involved in the morphological evolution of maize. They are also seeking to do the same from genes involved in the morphological evolution of different natural species of teosinte. 2) The second focus is a large-scale population genetics project (populational genomics) for maize and its relatives that involves labs at two other universities. They are attempting to define the population structure of maize and teosinte, estimate gene flow among subpopulations, and determine intraspecific "phylogenetic" relationships. Toward this goal, they have currently assayed 200 microsatellite loci across the genome, and they are trying to increase that number to 500. This large number of loci will provide the power needed to reconstruct independently the histories for different segments of the genome (i.e. individual chromosomes). His collaborators are Major Goodman at North Carolina State and Steve Kresovich at Cornell
University.

r) Michael Gribskov is from the San Diego Supercomputer center at UCSD. He is targeting functional genomics of plants, mostly Arabidopsis. He is currently studying two classes of enzymes, protein kinases and protein phosphatases. He is trying to get insertional knockouts from the Wisconsin facility of every one. So far, only three have a phenotype suggesting high functional redundancy. He seeks to generate databases, archive data, etc. to build a framework for information structure. Two major lines of questioning are 1)which groups of kinases have been elaborated among lineages and which ones common among plant species and can be traced through time using a phylogeny? 2) (with Mary Lou Guerno) what genes are involved in metal uptake (another few hundred genes) for example, genes involved in flavonoid biosynthetiss? He hopes to curate some 700 genes. He would like to address many common technology needs for performing operations for storing and displaying trees, and for doing analysis up and down trees. He is also involved with peripheral work with microarray data, EST assembly of possible splicing variants (mostly in animals, but could be also in plants), and how one metabolic pathway changes into another through protein evolution.

s) Neelima Sinha: Her lab is interested in leaf morphogenesis. Current work is using molecular and developmental genetics to understand leaf structure and diversity in leaf shape and degree of complexity . She commented on the NSF Workshop on Evolution Of Development (EvoDevo) and the Tree of Life held at Washington, DC. This workshop included representatives from plants, animals, and others. The workshop proposed that NSF should fund the generation of yacs, bacs, and EST tools for the next 100 genomes. This would be the takeoff point for further work in the evolution of development and in resolving ambiguities at critical nodes in the tree of life. She stressed that when the call comes out for proposals there should be some idea of the 40 to 60 organisms the plant group wants to focus on next and encouraged the group to begin strategizing.

t) Russ Chapman: Research in his laboratory over the years has focused on the Trentepohliales (subaerial green algae in the Ulvophyceae). The unusual cluster of ultrastructural features of these algae makes them unique. He uses RNA sequence data and other molecular data to understand relationships among the algal lineages, and the land plant - green algal relationships. The ultrastructure and biochemistry of Trentepohliales united the group with 4 of the 5 major green algal lineages (sensu Stewart and Mattox), a strange situation indeed! The Trentepohliales are in the lineage that did NOT give rise to land plants (i.e., they are in the Chlorophyta lineage not the Streptophyta lineage), but the Trentepohliales have a phragmoplast-mediated cell division. Some of the Charophycean algae closest to land plants (the Coleochaetales, Charales, and Zygnematales) have a phragmoplast-type cell division similar to that of the land plants. The presence of a phragmoplast-mediated cell division in both the Chlorophyta and the Streptophyta lineages indicates that the complicated fundamental cellular process of phragmoplast-mediated cytokinesis may have evolved in parallel. Homology or non-homology of the phragmoplast-mediated cytokinesis in the Trentepohliales vis-a-vis the charophycean green algae may be demonstrated by comparative analysis of genes associated with the cytolokinetic process.
Chapman Lab web site:

u) Bob Kuzoff: "I am completing a Katherine Esau postdoctoral fellowship at the University of California at Davis, in the lab of Charles Gasser, and will be starting my own lab this fall at the University of Georgia. My graduate work employed phylogenetic and classical microscopy approaches to explore floral diversification, especially architectural changes associated with the evolution of ovary position. As a postdoc I've endeavored to learn and employ molecular-genetic and bioinformatic techniques to extend my understanding of reproductive diversification in flowering plants. Presently, I am exploring the molecular-genetic basis of developmental transformations among ovules associated with shifts between anatropy and orthotropy, on the one hand, and bitegmy and unitegmy, on the other. If we take the common ancestor of angiosperms to be bitegmic and anatropous, then recent phylogenetic findings suggest that there have been at least 14 transitions from inverted to upright ovules and at least eight transitions from bitegmy to unitegmy across angiosperm phylogeny. Through cDNA library screening and EST database searches, I have obtained numerous orthologs of genes regulating these aspects of ovule development in Arabidopsis. I am comparing expression patterns for these orthologs in diverse angiosperms and plan to transform them back into Arabidopsis to see if they complement corresponding mutants in this species. My lab at the University of Georgia will continue working in this vein as well as obtaining and characterizing orthologs
of additional genes relevant to other aspects of ovule and gynoecial development."

v) Mel Oliver: His lab has been interested in understanding the underlying mechanisms of desiccation tolerance in plants. Much of his work has centered on the effects of desiccation and rehydration on gene expression in Tortula ruralis (a moss) and Sporobolus stapfianus (a grass). They are currently trying to pinpoint the genes that are central to all mechanisms of desiccation tolerance and those that are unique to individual systems. By doing this they hope, among other things, to track the evolution of this trait among all plants. His group has collaborated with an Australian group working with Sporobolus stapfianus. In his studies he is employing the tools of comparative genomics including the use of cDNA microarrays for expression profiling They are also in the functionality of those genes and in collaboration with David Cove Ralph Quatrano and others is trying to develop a homologous recombination system with Tortula. They have just begun their microarray analysis and have developed a small EST library. Their work includes biochemistry, phylogenetics, and genomics and accumulates data on functionality of proteins involved in the reassembly and repair of membranes. They have also, in conjunction with Andrew Wood and Doug Gage, written an NSF proposal dealing with comparative proteomics of the chloroplast, specifically a comparison of Tortula and Arabidopsis chloroplast protein compliments.

w) Dennis Wall (summarizing the Mishler Lab as a whole): The Mishler lab group is united under the common interest of phylogenetic systematics and comparative methods. Projects in the lab currently include phylogeny and biogeography of tropical mosses, California ferns, and medicinal plants of Polynesia. We have a long-standing interest in the phylogeny of the green plants in general and mosses in particular, and gather data at both the molecular and morphological levels. We are also interested in the theory of how to analyze these data in combined, large-scale fashion, and how to interpret the resulting trees evolutionarily (e.g., the relationship between evolution and development of leaves and peristomes, and the evolution of protein genes including codon usage bias). Several members of the lab work under an NSF PEET grant on the tropical moss family Calymperaceae and divide their time between the molecular lab and the field. New lab interests include incorporating genomics data in plant phylogeny.

LUNCH

1:28 pm
4) Scientific Issues:

Introduction: Brent Mishler commented on the diversity of the group of plant biologists here assembled and the goals of Deep Gene (what benefits the group activities should provide for the phylogeneticists and the developmentalist/functionalists). We hope that some members of this RCN will meet with members from the Deep Time RCN at the Botanical Society of America meetings in New Mexico in August 2001. He noted that the methodologies using genomics to build good phylogenies are applicable to many types of problems from resolving deep branches to looking at recent diversification events to understanding the evolution of function and form. He reviewed the utility of good phylogenies for those interested in functions and development. He first illustrated the idea that by doing sister group comparisons it is possible to reduce variance and narrow down predictions relating to the trait. This is done by comparing closest relatives that show a change in function (i.e. one taxon has the phenotype and one does not). This approach utilizes phylogenetics to set the stage for asking questions about the basis for the presence or absence of the function. In addition to sister group comparisons, ancestral-descendant comparisons would allow you to track the evolution of a trait from an ancestral state to its present complexity and find out what the genes were doing in the less complex systems. This approach would be more informative than trying to understand the basic elements of a highly derived complex system by studying complex examples of it. In addition to using the comparative methods approach to do comparative genomics, genomics can help to build better phylogenies. When we understand the genomics better we may be able to develop a better understanding of the evolutionary process.

Discussion plan: start with general ideas and move on to more specific ones.

COMMENT (Steve Tanksely): It has been said that phylogenies are more reticulate and less bifurcating in general than expected (in a recent microbial genomics review), so how does this effect phylogeny reconstruction since it assumes bifurcation?

COMMENT (Yin-Long Qiu): Most of us working on molecular phylogenies have assumed that the phylogenies are mostly bifurcating, but the distinction between prokaryote and eukaryote systems is significant. And for bacteria the reticulation is more of a problem but for eukaryotes there is little concern about reticulation for big questions in the Streptophyta (perhaps more concern at the shallower levels).

COMMENT (Brent Mishler): In a sense the phylogeny is a self-correcting system in that reticulation might lead to a "bush" but it wouldn't give you misleading result. In such a case where there is no strong signal the result would be a lack of resolution. Also, even for bacteria there are pretty good trees, which couldn't be developed if there were too much reticulation, implying that vertical transmission is giving an appreciably strong signal.

COMMENT (Chuck Delwiche): The precise measure of horizontal gene transfer in prokaryotes is unclear. Rosie Redfield argues against too much horizontal transfer but Doolittle says there is a lot (it is the dominant factor). Even if horizontal transfer is very common you can still construct an organismal phylogeny. Horizontal gene transfer is a complication but it does not preclude phylogeny reconstruction.

COMMENT (Steve Tanksley): If horizontal transfer is common in a lineage to the point that every gene has a different true evolutionary history you can still make a phylogeny if you have enough genes available.

COMMENT (Yin-Long Qiu): For the plants so far the horizontal transfer doesn't seem to be a problem. We are getting similar stories from different genes that are certainly vertically transferred.

COMMENT (Brent Mishler): Thus, it would be good to have a bigger suite of genes and nuclear markers to be compared.

COMMENT (Daphne Preuss): Similarly, the taxon selection available is an important factor. The group should come up with a list of "critical taxa." We should try to avoid a one-question-at-a-time approach and get the most data for the money.

COMMENT (Brent Mishler): Web site development is one of the goals. The Deep Green project did generate a key taxa list but we need a smaller set now for focus on taxa for genomic sequencing.

COMMENT (Mel Oliver): We need a basis for key taxon selection.

COMMENT (Claude DePamphilis): Other features could help select key taxa for the core list of taxa.

COMMENT (Steve Tanksley): When NSF first came out with the genomic sequencing programs, Craig Venter was interested and submitted a major proposal that was turned down in part because of question about the taxon selection for the project and help from this group for a well-justified list of 50 plant taxa.

GENERAL DISCUSSION: of EST vs. genomic sequencing; cost vs. data; prices are going down a lot; ESTs are quick and less expensive but genomic approaches provide more information which ultimately will be needed; need to consider size of genome, coverage.

COMMENT (Brent Mishler): Perhaps it is time to discuss the NSF workshops on a large scale funding for the Tree of Life - a post-genomics project (meetings at Austin, Davis and Yale). Terry Yates has mentioned a possible $100 million/year competition with several types of groups including a single national center like the ecological center at UCSB plus center-oriented grants (for museums etc.). Another component would be group oriented like Deep Green and Deep Gene, and clearly we should attempt to secure one of those grants if they occur. The final component will be to individual researchers. The biocomplexity program proposals are due in March. We could submit one for a select group of (10?) taxa to do chloroplast genomes and one for mitochondrial genomes for the same taxa. There should be no problem with two proposals if we avoid redundancy or inefficient use of the funding. They want to fund about 11 of these and we could ask for two or three. Of course, they must not be too dependent on each other and not too redundant. The first year this program was microbial genetics. This year is very focused in terms of what they wanted to fund via the biocomplexity program.

COMMENT (Mike Sanderson): Another approach is to choose 20 most difficult plant phylogenetic questions.

COMMENT (Yin-Long Qiu): We really want to know what happened in genomic evolution and thus the mitochondrial genome evolution is particularly appropriate and potentially useful. Some mitochondrial genomes are pretty stable and others (like Chlamydomonas) are very different.

COMMENT (Brent Mishler): If you are interested in the process of diversification and radiation, then recent groups (as opposed to deep lineages) would give you a better chance to explain the processes.

COMMENT (Daphne Preuss): Arabidopsis relatives would be very fine targets with CHPs. There are about close 40-60 relatives.

COMMENT (Mel Oliver): Are there any desiccation tolerant relatives?

COMMENT (Bob Kuzoff): It would be good to have things worked out for the close relatives and then move out to other plants not so closely related.

COMMENT (Steven Tanksley): Phylogenetics and function are two very different trajectories that use different genes for different questions. For this you need a very good nuclear gene database. It would be nice to have some closely related plants. Nuclear based datasets would allow a great deal of other work on nuclear genes.

COMMENT (Michael Zanis): Taxon sampling should include not only a bunch that span the tree but perhaps a few that are closely related to allow comparisons.

COMMENT (Toby Kellogg): How much information can you get from ESTs to answer some of the questions we have been discussing (a comparable role for a set of genes)?

REPLY (Steve Tanksley): There are some good questions that can be attacked via ESTs (examples in fruit development).

COMMENT (Neelima Sinha): We should think big because the questions that we want to answer over the next 10 years will require ambitious projects and a larger number of taxa--we should choose 50-60.

COMMENT (Paul Wolf): I agree with Michael Sanderson's suggestion of focusing on some key questions. We first need to ask, what are the questions and then where do we sample and what techniques do we employ?

COMMENT (Steve Tanksley): The phylogenetic-based taxon sampling is important, it allows us to follow a gene across lineages.

COMMENT (Brent Mishler): These are not mutually exclusive goals. It is clear that specific questions would require different taxon sampling. We already have a Deep Green list of 100-200 key taxa and could choose 20 key taxa from those. These taxa would not be appropriate for all questions. This is kind of a community service, a lesson for plant genomics. Pick 20-30 select genomes and let people see what they would be doing with them. These are two different research styles.

COMMENT (Chuck Delwiche): I agree these approaches are not mutually exclusive. A stratified sampling technique might be helpful, as well as 20 crucial questions we'd like to address. There might a group of about 60 taxa that would be an excellent starting point even for those questions that require other taxa as well. It is not a sequencing service to provide information on taxa just because they are key taxa. For each specific question, would find significant overlap, and simply provide the most useful resources for near and long time results.

GENERAL DISCUSSION: We don't even know all the interesting questions are yet. There are interesting questions we can ask right know. Deep Green did a lot of good things, but guiding the community to focus on selected taxa was a failure despite a lot of effort devoted to generating the lists. We need a list of organisms to help provide a phylogenetic framework. Since the costs are coming down and the capacity has been developed for the major sequencing projects, it may not be too important to worry about the selection of the taxa since we may be able to add taxa pretty easily. The development of analytical methods to handle the data generated will become very important. The computational problems are in some cases very complex. We don't have all the analytical techniques we need to generate the new and better phylogeny even if all the data were given to us. We will need prototype analysis methods and visualization methods.

[break]

3:11 pm
5) Business meeting focused on specific activities for the RCN grant this year

1) Brent Mishler: Discussion of the RCN grant operational details:
a) 3 workshops per year (2 in association with a professional meeting and one as a "retreat"):
i) this meeting
ii) Albuquerque, NM - 2nd week of August
Bot. Soc. Amer.
Biocomplexity groups
Deep Time group
Doug Soltis has organized an appropriate symposium
workshop on genomics techniques available
workshop on the "Exemplar Taxa" selection
iii) September Washington, D.C. Plant Genome project PIs will meet in a hotel in Ballston (satellite meeting? at Arley or some similar place) Liz Zimmer will host. Possible workshop topic: Tools for Phylogenetic Study.
iv) Phycology Society of America workshop in Estes Park, CO low cost informational meeting (perhaps a lunch)
b) exchange of graduate students (6 per year at 3 months each)open to the students of the collaborators
c) summer interns for undergrads 12 per year open to the students of the collaborators
d) K-12 outreach for teacher 1 working meeting per year (rotate around because we can't provide transportation for a large number of participants)
e) web support (grad student)
f) Steering Committee of 6 plus the PI (3 phylogeneticists and 3 genomists, initially from the original collaborators that wrote the proposal, to be rotated next year). Suggested slate: Steve Tanksley, Chris Sommerville, Daphe Preuss, Kathy Pryer, Toby Kellogg, Yin-Long Qiu unanimously approved
g) graduate student representative: Jeff Lewandowski
h) membership/collaborators: should be expanded
i) there is a need to list major groups of taxa on which various
investigators are working (especially important for graduate students who are trying to choose their dissertation problem so they don't get scooped etc.) Criteria for choosing taxa are on the Deep Green web page but need to be revised.

COMMENT (Toby Kellogg): The discussion has focused on developing better phylogenies through organellar genome studies but it should be noted that the need for nuclear genes and genomes studies is key for developmental biologists. Also, antibodies or proteomic techniques that are important for developmental studies are of interest and importance to the developmental biologists. Having EST sets available would help. The in situ hybridization approach is not efficient (too slow) unless you already know the gene is one that is interesting.

COMMENT (Chris Sommerville): It is important to develop a large set of ETS for taxa in which were are interested so we can develop oligos needed to look for genes of interest in any other species.

COMMENT (Brent Mishler): Early on we need a set of ETS and need to get a grant to develop such a set, but remember PIs need to submit proposal - this group doesn't submit proposals.

COMMENT (Chris Sommerville): For the Arabidopsis group they had workshops and got NSF to issue RFPs that reflected the recommendations of the community. This way, somebody will submit the proposals and the work will get done.

COMMENT (Brent Mishler): Agreed, we need to help guide NSF proactively, but we also need to be applying for grants as they come up.

COMMENT (Paul Wolf): Are groups in other countries who are doing this kind of thing? We do want to reduce overlap and duplication. There followed a general discussion of what is going on where. Members of this group should share information that they may have about who is doing what etc.

Concluding remarks by Brent Mishler: We have had a productive one day meeting and need to continue to share information and keep thinking about what we can and should be doing.