What sorts of categories should we have?

Off the top of my head:

• alignment
• macroevolution
• software
• trait evolution
• fossils
• papers

This is nice, Erick. Thanks for setting this up.

With regard to topics, perhaps it’d be nice to have a topics on:

• molecular evolution
• open positions: postdocs/studentships/jobs
• workshops
• phylogeography

The site looks great!

Does the software support nested groups? This might help with organization, administration, filtering topic feeds, etc.

e.g.

Community

• general discussion
• position/workshop announcements
• talks, visiting scholars, etc.
• software releases

Research

• models
• alignment
• trait evolution
• divergence dating
• …
• datasets
• code/scripts
• paper discussions
• troubleshooting phylo issues

@mlandis I like that suggestion, but the categories are “flat”, disjoint, and dead-simple. By design, I’m sure.

After a chat with @trvrb I don’t think that we can have both “type” and “subject” categories. I advocate the latter, as people are more likely to be interested in the majority of posts in a given subject than all papers, for example.

• alignment
• data (incl. NGS)
• macroevolution
• molecular evolution (sequence models, etc)
• software (the actual practice of using software to crunch data and visualize)
• traits (incl. morphology, also phylogenomics)
• dating & fossils
• phylogeography
• tree reconciliation (gene-tree vs species-tree, etc)
• miscellany (noting that we already have meta)
• workshops & meetings
• positions: doctoral
• positions: postdoc
• positions: jobs (split into “faculty” and “other”?)

I don’t want to make a “Bayesian” category because that’s pretty cross topic, but what about a “tree posterior” category? Or a “confidence” category?

How about popgen and genetics categories (with the implicit idea that people will be discussing the parts of these fields that touch phylo?)

Other ideas:

• measurably evolving populations
• taxonomy (and species delimitation)
• math (phylo theory)

My sense is that fewer categories are better, but we want everyone to have a place to post.

Popularity will dictate category count to some degree. A very popular forum might demand more granularity (so if I’m looking for postdoc announcements I don’t have to visually filter grad student positions). However, if there’s only one post a month, it would be better to have a single ‘positions’ category.

I like the proposed ideas. My take (trying to combine a bit and shorten names):

Community

• Talks, workshops and meetings
• Positions

Research

• Software (including releases)
• Alignment
• Traits (including macroevolution)
• Sequences (molecular evolution)
• Dates (including fossils and measurably evolving populations)
• Geography (I think this is too popular to be combined with Traits but I don’t know)
• Data (what would be here besides NGS that isn’t covered by the above data types?)
• Species trees
• Pop gen (including coalescent)

Other

• Meta

This is fantastic. I like the combinations.

I suggest “Events” as a name.

It all depends on who shows, right? I suggest we start with a minimal set and add more categories as folks show. It’s easier to add then remove. With this in mind, I would suggest moving Geography into Traits. Perhaps also get rid of pop gen and data?

The flipside is that we don’t want to make people unenthused because their fave category is missing.

I dig it. My revisions:

Community

• Events
• Positions

Research

• Software (including releases)
• Alignment
• Traits (including macroevolution and geography)
• Sequences (molecular evolution)
• Dates (including fossils and measurably evolving populations)
• Shape (not sure on name, maybe Structure or Topology. This would be tree topology / structure issues, that includes the coalescent and gene tree / species tree)

Other

• Meta

Excellent. I’m going to vote for “Topology”. Others-- approve? If people are happy I will set them up.

(cc @mlandis @trayc7 @phylorich @josephwb)

+1 seems like a good starting point; presumably they can split and merge over time if one gets too big.

It might be worth some way of distinguishing between specific requests for help with particular pieces of software (program X won’t run on my data – help) with more general topics (what are limitations of model Y?); this is one reason why I don’t follow r-sig-phylo very closely as it’s hard to spot the more generally interesting topics at a glance.

Rich-- this is a very good point. Perhaps we should have the Troubleshooting topic that @mlandis suggested and be strict about keeping all such discussions there? We want this to be interesting and useful for methods developers.

Seems like a great idea to branch that off.

I have added the categories along with draft definitions. It turns out that you can have nested categories, but I think that leaving it flat at this stage would be a good idea.

I have also added the following bullet point to the “you are new at posting” guidlines:

• Is this post a request for help with a software package? This is a forum for discussion, not for debugging. We do have a category for discussing specific analyses, which is Troubleshooting, but if your question is about how to run a specific package we suggest that you address the user group for that package.

Please help me make that sound less passive aggressive.

Hey folks! My sense is still that there are too many categories - people are going to be trying to figure out what this is all about, and keeping it simple to start would help. I am tempted to combine all of the theory into one category for now, and then separate it if and when the number of posts increases to the point that separation makes sense. This also might help with the distinction between “discussion about methods here” and “troubleshooting analyses here”.

Rather than collapse all of the theory into one category, I’d rather have more posts! I think that as soon as there are enough topics to push down the category definitions, it will give people something to say “oh, I see what this is about”, rather than “WTF”.

I think I might have an announcements category where people can put software, events, etc. So they all go one place. Otherwise some of these might get a little noisy? Dont’ really have a strong opinion though

So we do have a “Software” category and an “Events” category. Would you suggest merging those into an “Announcements” category?

If the software one is just announcements about new software then I would do that. I imagine though people might want to talk about running software (as opposed to coding which is another thing) as well though so I might leave software and just have announcements about software go in announcements. I would put all announcements in one topic (workshops, open positions, software) and people can just have [SOFTWARE] in the topic. That said, again, I am open to whatever, just a thought.

I’m happy to go for “Announcements.”

Do folks have any other proposed merges? I’m hesitant to move all theory into a category because I think that there should be more communication between theoreticians and practitioners. But I’d be happy to be overruled by others.

How about an EXAMPLES category with both good and bad examples from the published literature. I’ve collected quite a few papers that are great examples because they have fantastic data sets publicly shared in TREEBASE, others that are great because they show a new method, etc. Plus I have some stunningly bad examples, along with my analysis of what went wrong.

If you have a question about a specific software package (“XXX won’t run on my machine”, “What does the YYY parameter do?”), please try the user group for that package or contacting the software authors. We have a category for discussing specific analyses, which is Troubleshooting.

I think the Topology description should include phylogenetic networks, eg for hybridization.

I am not very happy about Sequences and Alignments, and suggest instead

Molecular evolution: Nucleotide and amino acid sequences, including polyploidy, synteny, protein structure, and modeling their evolution.

and

Preparing data: The theory and practice of sequence alignment, tests for recombination, paralogy, etc, using morphological characters, what to do with ‘too much’ data.

I don’t know where I’d post about species delimitation.

Nomenclature (or Taxonomy)

Thanks everyone for their input here!

That sounds interesting and useful, but I think that examples should be categorized with what went wrong-- e.g. did the researchers mis-estimate branch length support?

I like this. I’m enjoying having categories be a single word, though, so how about just “Data”?

@tmkeesey I’ll be happy to add this if people post in it! I suggested a taxonomy category before but it was dropped.

Data is fine. Maybe Taxa for taxonomy and species delimitation.

mmmm… I thought in something concerning “organismal” issues ( something more broader that would include taxonomy, and its thechnicalities; species concepts; species delimitation…) But I could not think in a better name than “Organisms”

Any Ideas?

I would like to see a place for discussion of low-level tools for gathering data sets and getting the data ready for analyses. This work is the bane of my existence when I go outside the Human Immunodeficiency Virus database to gather data sets.

In the genomics/proteomics era, the problem is the flood of data and the sparse annotation of that data. And when the data is annotated there is a lack of consistency across data sets. Below species level, terms like “race”, “strain”, “isolate”, “type”, “serotype”, “Subtype” and so on, are frequently critical in the study of viruses, bacteria, plants, animals or other organisms. For some data sets we may want geographic location or other data. Many tools will easily pull “ORGANISM” or “Genus-Species” from the database, but have trouble getting the other data, largely because the data is inconsistently annotated to begin with.

It is clearly a pipe dream to wish that NCBI/GenBank or other databases will provide all the information for all organisms and all types of analyses of any given group of organisms. But there are many tools already developed, and more being built each day, to assist researchers in gathering and organizing their data sets.

For one example, today I just learned of BLAST-EXPLORER from http://www.phylogeny.fr which was published in: Dereeper et al. BMC Evolutionary Biology 2010, 10:8 http://www.biomedcentral.com/1471-2148/10/8 http://www.ncbi.nlm.nih.gov/pmc/articles/PMC2821324/ Which from the publication looks like it will work great for many vertebrate or plant data sets, but failed badly during my “test drive” because I work with viruses and bacteria.

After gathering and aligning a set of sequences we can still be faced with trivial (from a science view) but annoying issues of getting the data into the right format for analyses. For example our favorite multiple sequence alignment tool may like to input or output in FASTA format, but we need a different format (NEXUS, SLX, GDE etc) for input to another analysis tool.

We clearly don’t want to lump this type of discussion in with other topics which are supposed to be about the real science of phylogenetics. And perhaps we don’t want to have such discussions here in the PHYLOBABBLE site at all. But I would find it very handy to have a place to share ideas and 'tips and tricks" of the trade.

Thanks for your thoughts, everyone.

Here are my proposals:

• A new “Preparation” category that has everything related to gathering data and readying it for
• A new “Taxa” category that has organisms, taxonomy, etc

Feel free to make further suggestions. Or you can “like” this post to show your support.

OK, well, I didn’t do what I promised, but I think it’s better.

Sequence alignment and tests for paralogy, etc, are quite nontrivial and so I don’t think that they belong with how to download lots of data. Thus I have changed “Alignment” to “Homology”.

To handle the requests of the Data-folks, I have started a “Practicalities” category, which also includes the visualization of trees, etc.

There is also a new “Taxa” category.