[Note: these are rough minutes taken down during the meeting itself. They are intended for archival purposes; please excuse the rough edges -- we would appreciate getting feedback from anyone about errors or omissions in this document.]Participants Present:
|Mark Buchheim (organizer),||email@example.com|
In discussion--Many taxa talked about in large groups of convenience
Chuck Delwiche, molecular data not strong enough yet to give strength to different names to groups.
What is the life history of Codialales--heteromorphic alternation of generations
Mark--does anyone know about definitive life histories in Chaetophorales? None were aware of any recent work on chaetophoralean life histories
Matt Cimino: Nomenclature of Coleochaetales--he's looking at 22 species altogether. What are good species, what needs to be republished, conserved, etc. Working at generative level. He has found about 6 additional species. Two species that don't really fit into any others--have divergent sequences. Some that are morphologically different in subtle ways--has 15 or 16 in culture. He is working on 18S data set. He is hoping in the spring to get in the field to culture some more Chaetosporidium. Takes about 3 months to get enough biomass to sequence, probably won't have ready for IBC. Hopes to have plastid data set, maybe mitochondrial data set, for IBC. Chuck Delwiche says will have 18S, rbcL. For the Charophyceae will have at least atpB plus 18S, rbcL, maybe even another gene. Also wants to look at cox3. Hope to have written up in press before IBC. Matt Cimino hopes to get some more species from field.
Ken Karol: incongruence issue between rbcL and 18S phylogeny in the charophycean realm. Huge difference in tree topologies. He has been looking at 18S data, and quality is not particularly good. Also, looking at alignment issues in 18S molecule. Found some secondary structures on web to use as starting point for aligning stems, loops, etc. Looked at areas that couldn't be aligned by eye, folded on MFold. He is starting to make a new alignment. Getting through 2/3 of gene--didn't look bad at all in terms of consistency with rbcL. Stem changes compensatory, so weighted by 1/2--then gets topology nearly identical. Increasing bryophyte taxa in his data set now, so must rework alignments and refolding as needed. Problems--stems and loops change at bases--must decide whether to weight. Bulges sometimes in stems. Dynamic of base pairing can make bulges become pair, which will make different bulges become pair.
Chuck Delwiche --thinks that the functional homolgy portion of gene can drift.
Rick Zechman--"functional" homology implies non-homology--this is why he doesn't like using secondary structure for alignments.
Ken--molecule is functional.
Chuck Delwiche-function of molecule is selective.
Rick Zechman-how do you decide what is homologous? He would argue that the sequence similarity is homologous and different structures are selected.
Marvin Fawley --sequence evolved, but structure conserved.
Ken--single base bulge on stem
Chuck Delwiche--bulge is selected for. This is a fundamental problem for 18S. In primary sequences, bases are homologous, but selection is conserving a feature that doesn't carry information needed for phylogenetic reconstruction.
Chuck Delwiche--hope that number of positions that change like this is small. It does mean that you have to be careful with alignment because secondary structure features can confuse.
Ken Karol--but secondary structure can be very helpful.
Chuck Delwiche--alignments that Ken has been making are equally as plausible as others, but the topologies are very different. Not unique to ribosomal data, also protein coding data.
xx - xxx
In this example, what is the true the homology? How do you define?
Marvin Fawley--is it always clear which base is bulging? Can you usually tell?
Ken Karol-usually. The published structures have nice big loops, but when you fold there could be some differences.
Chuck Delwiche--also, there's a known 4 nucleotide sequence that forms a micro-helix, a stack, in RNA secondary structure characters. Rules that go beyond stems and loops--we need to take advantage of these.
Marvin Fawley--is there is a temperature dependence for all this?
Chuck Delwiche--there is--thermophilic ones in sequences with high GC content.
Marvin Fawley-proteins involved in stabilizing.
Chuck Delwiche--initial guess that the Charophyceae had lost one of ribosomal proteins, but now looks like a matter of really careful alignment, looking at every indel and asking, why are you putting indel in? Can it be justified?
Rick McCourt-hypothetical example--if aligning very divergent taxa. If have a region of 5-10 bases that are exactly identical, but sequence shifted to accommodate structure--do you align with sequence or to accommodate structure?
Chuck Delwiche--not hypothetical--see drawing--thinks all through point mutations. Would not be inserting indels
Marvin Fawley--problem is in inserting gaps and then getting rid of similarity information, unless code insertions, can lead to great alignment problems. Maybe not be able to easily align.
Chuck Delwiche--characters that using are nucleotides--following changes in nucleotides. So, need to focus on--in this column of amino acids, is there a good reason to believe they are all descended from the same nucleotide ancestors?
Marvin Fawley--when looking at sequences of organisms that are very different from each other, especially different?
Ken Karol--not all regions are alignable by secondary structure.
Chuck Delwiche--if not alignable--then don't use. Can try to code in a different way. Character types other than amino acids and nucleotide are more difficult to work with. Had data set with nice indels--tried coding separately. This is serious problem and doesn't think there is a good solution. Have to be careful, especially with ribosomal data--also recognize that homoplasy happens.
Mark Buchheim--Is there any evidence that secondary structure--in certain ecological areas, is it distinct?
Ken Karol--someone has looked at in cacti.
Marvin Fawley--many cacti are exposed to both low and high temperature extremes.
Chuck Delwiche--are you asking about secondary features that are different?
Mark Buchheim--yes, cold vs hot climate types.
Marvin Fawley--seasonally exposed stuff-will grow fine in room temps.
--divergent conversations here--
Marvin Fawley: If there is a selection on certain structures depending on hot or cold, then there must be something that functions to stabilize.
Chuck Delwiche--if culture growing under wide range of temperatures that seem all the same, might try cloning.
Marvin Fawley--routinely examines with rflps and they seem the same. Thinks it does bring up interesting questions.
Mark also has a "cold" chlamy that grows well at room temperature. Collected from cold habitats--"cold" Chloromonas species don't have pyrenoids--A group of plant physiologists were convinced that pyrenoids are necessary for carbonic anhydrase--he sent them a culture, but they couldn't find a pyrenoid. He is a little concerned about what is going on with cold water creatures, but they are so similar at the 18S level.
Chuck Delwiche-wonders if there are sequence motifs recognizable in rbcL that correspond to pyrenoids.
Break-- Talk during "break" : ranking of groups--Chuck Delwiche thinks Brent will want ranklessness for IBC--Chuck Delwiche doesn't think it is a problem.
Europeans have not accepted the Charophyceae sensu Mattox and Stewart.
Rick McCourt--rbcL wins because has more informative sites, but bootstrap support is higher for 18S.
Rick Zechman--evidence for saturation in rbcL.
Rick McCourt--how do you measure saturation?
Rick Zechman --transistion/transversion ratios.
Chuck Delwiche--easiest way is to calculate maximum likelihood distances with most complex model for data--if distances get above one (taking into account invariable sites--model must be invariant sites model) then there is saturation.
Rick Zechman-why does this translate to saturation?
Chuck Delwiche--curve observed distance vs real where if you have perfect data, would fall on line, but there is a fall-off so 25% of characters match at infinity. Distance models basically take into account multiple substitutions expected and so move data back toward the line, so distances go toward infinity as get more and more substitutions, so if get distance of one, then one implies substitution at every site.
Rick McCourt-Louise Lewis did something interesting with bryophytes--taking out 3rd position.
Marvin Fawley-analyzing is important because saturation does occur, but can look like saturation and not be--that is where resolution is.
Rick McCourt --Louise is talking about bryophytes and charophytes?
There is fuzziness at the base of the clade--is Mesostigma at the base of the clade? Supported but not well-resolved in rbcL. Klebsormidales interesting and he plans to work on them. Will models be homologous or analogous?
Marvin Fawley? Have you found any way to screen easily?
Chuck Delwiche--wants to do environmental study, extract DNA from scum and try to design primers specific to Charophyceae, but not too specific, amplify, clone, then use hybridization to find those organisms close to the Charophyceae in mats at vernal pools. If can see the organisms, can use fluorescent probes.
Marvin Fawley--must recognize that there are pitfalls along the way.
Chuck Delwiche-could be selective amplification problem. But if screen enough clones, should find interesting ones and will be better than individually culturing and then screening cultures that may not even be interesting. Primers not yet designed.
Marvin Fawley--cultures first then screens with specific primers and conserved restriction sites. With Chuck Delwiche's approach can see how much might be out there first.
Chuck Delwiche--wants to see how much diversity is out there. Difficult with morphologically very similar organisms.
Marvin Fawley--looking that this type of thing with coccoids--easier to screen from cultures, but may only be sampling a fraction of the diversity there.
Chuck Delwiche--environmental sampling with capture-capture type statistics. This is a real advantage is finding things that can be cultured.
Marvin Fawley: Prasinophyte exemplar taxa identified at Breckenridge meeting (see meeting notes on web from august 95)--fragmented data. Table of exemplar table has stood up well, even though we have more to do. Today-- exemplar taxa expanded --filled out well for rbcL and 18S, mainly because much has been put on GenBank very recently from Melkonian. A few gaps in rbcL--some in 18S (Resultor, Pedinomonas, Scourfieldia, Tasmanites). This data matrix is from website that Marvin Fawley filled in some. There is a real Pedinomonas, unlike the earlier Pedinomonas that is actually a Chlorarachniophyte. Good progress is being made, but a lot of it is outside the exemplar taxa identified at Breckenridge.
Neighbor-joining tree done this week--alignment not good enough for parsimony. Tree does have good structure. Upper part of tree extremely well-supported, including all higher "chlorphyta." Tetraselmis falls out there rather than basal. Tetraselmis, not trebuoxiophytes, not prasinophytes--thinks Rick Zechman should consider including with "ulvophytes" in molecular data.
Concept of Prasinophyceae has changed dramatically since mid-80s when they were "scaly flagellates." Micromonas is naked, Mantoniella has scales--very closely related by molecular data. Protein data also show this similarity. Now some coccoids are considered and some non-scaled organisms such as Micromonas and coccoid CCMPs. Now neither scales nor flagella are necessary. Back to Mattox and Stewarts--"paraphyletic" (although they didn't use the term).
CCMPs are coccoids which Marvin Fawley has worked on recently.
Pseudoscourfieldia. Pycnococcus come out together with rbcl, but not Nephroselmis. Marvin Fawley's redone Pseudoscourfieldia also comes out with Pycnoccous and not with Nephroselmis as in previous studies. CCMP 1198 is probably another isolate of Pycnococcus. The previous Pseudoscourfieldia studies had a misidentified isolate of Nephroselmis (CCMP 717?)
Mamiellales including the coccoid Ostreococcus and Mantoniella, with Micromonas pusilla.
CCMP clade of coccoids. There is a CCMP 1205 done by Bob Andersen that comes out in different place.
Mesostigma not available.
Interesting is some of other characters. Similarity of Micromonas and Mantoniella he had established earlier with pigment proteins.
Separate lineages with prasinoxanthin and similar light harvesting complexes, which no others have.
Minor pigment suites similar in many small groups and different in others. This is unususal in higher groups. So different lineages are held together by these minor pigments and some other tings.
CCMP 1205 seems to have something completely different.
Another tree done Thursday morning:
Mesostigma comes out with Streptophora. Rest of tree confused, but lineages hold up. Rooting problems maybe?
Chuck Delwiche--Bryopsis and Codium need to be looked at more closely.
Rick Zechman--they have a huge intron--15K base pairs. Seems to be only in Bryopsis and Codium.
Marvin Fawley--structure in terms of lineages, not where the individual taxa fall out, is very similar. Pedinomonas and Resultor coming out together. Lineages holding up, analysis needs to be worked on. Further work--still working on coccoids. Doing 18S, working on rbcL stuff now with Debra and Chapman lab.
Chuck Delwiche--such important organisms, really need a bunch of genes from these taxa. Really worthwhile to accumulate a DNA library so when resources are available, could work on other genes. Need a larger number of glaucocystophytes, need rbcL.
Marvin Fawley--Rooting problems, even more a problem with rbcL, the groups obviously very different, so how could not be saturated?
Chuck Delwiche--could they be completely saturated?
Marvin Fawley--not completely.
Matt Cimino-would gene order data be valuable?
Chuck Delwiche--are you saying for rooting problem or within the group?
Marvin Fawley--within....need a better outgroup.
Mark--what is rationality for using glaucocystophytes?
Marvin Fawley--came out more similar than anything else.
Chuck Delwiche--compelling set of data that say primary plastids are monophyletic and within this group the glaucocystophytes are appropriate outgroup taxa unless saturated.
Marvin Fawley--in his analyses, they always come out in charophytes, which is not good.
Chuck Delwiche--if the first split is between charophytes and other greens, then that is not that outrageous.
Marvin Fawley--no, but it does skew analysis.
Break for lunch
The Morphological DAM
Rick Zechman--deleted character 1--we know enough about green algae to know the value of these characters--for instance habitat, in one genus there can be three different habitats, so will not be informative for higher level relationships.
Deleted substrate attachment character 3--across all green algae not good (Chuck Delwiche--could code as multi-); presence or absence of contiguous vegetative cells--character #6 deleted --seemed artificial.
Mark Buchheim--this character was designed to differentiate Volvocales.
Rick Zechman--deleted this character, but incorporated it into a different Character #67--presence/absence of proximal septum--he incorporated this into character 66, presence/absence of transverse septum--changed to present, intercalating, present non-intercalating (for instance do Micromonas and Mantoniella have two septa?
Mark Buchheim--yes, but it is very difficult to tell if it is transverse or proximal in some taxa, as well--this needs more work. Do we need to worry more about what the drier plant people are doing?
Chuck Delwiche:--only for global analysis--will have to talk about certain characters then to be sure everyone is using same definitions.
Marvin Fawley--how far does this data set go?
Rick Zechman -those are the only characters he changed.
Chuck Delwiche--generally agrees with what Rick has done. But this is the problem with morphological data.
Marvin Fawley--global/morphological analysis--in between is a group of unicellular things that separate the two lineages of charophytes/higher plants and the basal lineages. Seem to be independently derived. Anything that has multicellularity has to be independently derived. Connecting with unicells. Why would we do a comparison where we know there can't be anything but homoplasy?
Rick McCourt--some of these things are in the Charophyceae.
Marvin Fawley--but to get to multicellular green algae, have to get to unicells.
Rick Zechman--maybe we shouldn't assume that we know nothing.
Rick McCourt--some of the characters can be defined superficially, and we can reject homology. Wouldn't throw in multicellularity.
Marvin Fawley--maybe there are better ways to define multicellularity.
Matt Cimino--redefine the characters so that the definitions lead to certain groups.
Rick Zechman and Marvin Fawley--this is a slippery slope!
Chuck Delwiche--this is why he is doing molecular systematics--morphological work is incredibly valuable work, but morphological cladistic analyses--with sufficiently clever character definition can make things turn out however you want. He prefers to do molecular and then map morphological characters. Molecular data is causing us to reinterpret morphological data. He doesn't believe that morphological data can give resolution where molecular data can't.
Marvin Fawley--doesn't that partly depend on the organisms and the level you are working at? In some groups morphological data is useful,
Chuck Delwiche--it is useful in determining high level relationships. He has completely rethought morphological characters since he's done molecular work and now can recognize synapomorphies that he previously thought were variable. Molecular trees seem completely reasonable, as reasonable as any hypotheses he had before. He is skeptical of the utility of morphological data for determining the branching order in this problem at our current state of knowledge.
Rick McCourt--this was reflected in Donohue's talk (in Baton Rouge) where he showed that probably right in the first place--the morphological data isn't worth much unless we examine it very closely. If we examine this matrix very closely, we might want to throw a lot more out. If you want a quick tree, need to do molecular work, because to really understand morphological characters would require a lot of, say, EM work.
Chuck Delwiche--we need to proceed and look carefully at a lot of morphological data, but you get more bang for the buck from molecular with morphological characters mapped on.
Marvin Fawley--sometimes it is difficult to see patterns until we have guides.
Chuck Delwiche--we're at a point where we have a new thing--molecular data. These data are giving us cause to examine morphological data again. L. Graham made a fairly convincing argument to look at parenchyma, which gave Coleochaete the nod as being the green alga most closely related to higher plants, but molecular data show that the parenchyma in Coleochaete is not homologous with higher plants. Instead, molecular data have raised new morphological questions about the arrangement of tissues.
Marvin Fawley--the whole thing that has to be avoided is searching for those cases that will support your molecular tree and ignoring those that don't. That can become a trap. When we think about where we started, we usually started with a hypothesis based on morphological data, and we would test that hypothesis. When we do a molecular tree, we need to test the morphological hypothesis, not attempt to support it.
Chuck Delwiche--look for congruence vs conflict in mapping morphological characters on molecular trees. When you look at morphological data sets, there is a constant conflict on what makes good characters. Chloroplast forms in the Zygnematales, for instance. Molecular data make chloroplast shape look important and filamentous state not as important.
Rick McCourt--do you include all these characters anyway and let others squeeze them out?
Chuck Delwiche--so can we phrase the problem in terms that one thing is that we should have uniform characters within OTUs? We are using exemplars, really, and not OTUs.
Matt Cimino--can we use the type?
Chuck Delwiche--that is a very taxonomic answer--doesn't work.
Rick Zechman--based on our current understanding of green plant phylogeny-if there is a question of whether a character is homologous, then it should be left in, but if they are clearly not homologous, then they should be left out.
Marvin Fawley--where do we draw the line?
Rick Zechman--if there is a question, include, if not, don't (such as habitat).
Chuck Delwiche--code the character, but then exclude it from the analysis.
Matt Cimino-try to include a discussion about each character.
Rick Zechman --but we don't want "click" analysis.
Chuck Delwiche--we need to leave tracks so people will know where the characters came from.
Rick Zechman--we are looking for congruence between different data sets, and so we need as good a data set as we can get.
Rick McCourt--I think, overall, we are worried about the quality of the morphological data set. If we go character by character to see if it is scorable legitimately...Where do we think higher plants came from, freshwater or marine? The modes of attachment are not similar
Chuck Delwiche--it becomes a question of coding.
Rick Zechman--what about character 5? Growth form: multicellular , unicell or coccoid, coenobic? What I have the most problem with is varying, general characters.
Chuck Delwiche--much better to divide into very detailed characters that can be coded with confidence, plus or minus.
Mark--I think we are going to have a transitional matrix.
Rick McCourt--If you have them in there, they come out mapped in an interesting way.
Rick Zechman-will be the repository if others will check the character states of the characters.
The group spent a little time looking at the GP-morph data set to see if it were correct. What follows is some of the changes and discussion:
Klebsormidium---no plasmodesmata. no parenchyma in Coleochaete,; Coleochaete is branched; some discussion about holocarpic vs heterocarpic, multiple sporulation/fission (score and delete???); type of sex--Mark-some of these seem to be guesses, some are not known to have sex; chlorplast shape--Rick Zechman added new state--multiple angular--Chuck Delwiche--chloroplast in Coleochaete should be parietal--but that is not a character state, so changed that and Klebsormidium to incomplete rings; changed pyrenoids in Klebsormidum to absent; but yes in Coleochaete for thylakoid membranes; number of flagella; papillae on vegetative cells?; crystalline cell wall--a Chlamy thing; specialized zoosporaniga--changed Coleochaete to yes; apical cell growth-changed Coleochaete to yes; Klebsormidium--everything to do with sex changed to question mark because sex is unknown--many things with Coleochaete and Klebsormidium changed in sex; Coleochaete changed to question mark for cuticle; independent sporophyte--this needs to be redefined so does alternation of generations with an embryo as part of definition
Rick Zechman added the characters: xylan in cell wall of gametophyte, present/absent; mannin in cell wall of gametophyte, present/absent; cellulose in cell wall of gametophyte; present/absent; habitat-independent siphoning; leucoplasts present or absent; codiolum sporophyte; cell wall calcification; gametaniga in siphonous; flattened flagellar apparatus.
Mark Buchheim--if Russ is going to give the overview of the molecular, who is going to do the overall alignment?
Break for dinner.
Resume in the Adam's Mark lobby.
Chuck Delwiche reports on Executive Committee meeting in October:
R. Chapman, P. Soltis, E. Zimmer, C. Delwiche, B. Mishler, met last fall to talk to the USDA to update them on the utility of the grant and get guidance for future directions. For instance, how can we maintain the benefits of the group?
Met with around a dozen representatives from the three groups: USDA, NSF, DOE; including Jim Estes of the Plant Genome Group, Jim Tavares, DOE, Karen Kindle, Jim Rodman, NSF, and the USDA program officer. They spent about an hour with the group and were really interested in what the group has accomplished and in future directions.
One thing that would potentially be useful is the USDA experiment stations--cooperative agreements," with regional study groups, but since not many researchers in the GPPRCG are funded by USDA, that is not practical. Other possibilities discussed--genomic funding, bioinformatics funding.
The greatest strengths of the group are in analyzing data and knowledge of organisms.
For the genomics projects, the rationale, the mission of the genomic programs, is to get information about economically informative plants. It is not financially possible to get the genomes of all economically important plants, so it is important to extend that knowledge to other economically important plants. Half to 2/3 of all agriculture involves about ten crop plants--corn, rice, soybeans, wheat--these 4 crops comprise about half of agriculture, the other half is comprised of about 500 plants--so called orphan-plants.
Therefore, need a good comparative context to extend knowledge to these orphan plants.
Chuck Delwiche was putting together a proposal for cDNA sequences for representative plants picked as keystone taxa. With big ticket genes, so 5,000 cDNAs from a small selection of plants. The idea behind the proposal is to ????? a large number of coding genes and help define priorities.
Model systems in economically important plants.
Comparative sequence analysis.
Knowledge and distributed information.
The LSU and Princeton meetings seemed to be really helpful in bringing people together.
What has not been dealt with is what to do with the green algae people. Avenues for funding to bring together the groups--small directed groups--to continue talking about what data we have and how to put it together. Should try to work through PSA meetings and have an extra day of directed meetings. Try to stay together for at least another 5 years. Try to get $10,000 or so a year to support the group.
NASA--astrobiology? Fundamental early backing--in developing of 02.
Biotic observatories--a new program starting soon. Field stations with people focusing on certain groups of organisms. Taking diversity to the ecological level. Simultaneous monitoring of places.
Private funding? Pharmaceutical company?
Should write a prospectus--what sort of meetings we would have and list of potential participants.
Format for book--editors Mishler, Doyle, & McCourt. Phylogeny 13 chapters. Deadline for Phylogeny, February 2000. For Evolution, April 2000. The product will include a CD Rom. There will be 3 or 4 coordinators.
At IBC, there will be six symposia, 8 speakers at each. Brent Mishler will give a summary of all plants. Russ will give the summary of all green algae. GPPRCG is sponsoring 10-15 graduate student travelers.
Target date--April 1 for abstracts.
Molecular data due date? Huelsenbeck is willing to help if he doesn't have to do it all by himself. Ken Karol will help. Molecular data should be submitted by the end of May or Mid-June, which will give time to have papers submitted. 18S that has already been published should be sent to Ken Karol as soon as possible. Also, send him alignments. Once the alignment is completed, as many people as possible can work on the analysis.
End at 9:30 pm.