The list of blogs being indexed by Postgenomic has risen sharply in the past couple of days, thanks to some fantastic work by Bill Hooker (of Open Reading Frame) - thanks Bill!

The new blogs still have to be checked and categorized, so one or two might disappear by the end of the week.

Postgenomic now clusters posts, memeorandum-style. This should give you a good overview of what the science blogosphere is currently talking about.

Occasionally the clustering algorithm is fooled by articles that share a large number of key words but aren't actually about the same thing. Over the next few weeks I'll keep an eye on the clusters and tweak things accordingly.

The Postgenomic code is now available via subversion from Google Code.

This is an "alpha" release in that I'm still in the process of removing Postgenomic-specific features: some of the code is designed to work with data from hundreds of blogs and isn't really suited to smaller installations. The documentation is also rather poor.

The code has been completely refactored to make it easier to adapt to your own needs, though the overall architecture is still the same (there's a pipeline written in Perl which downloads feeds, processes them and writes to the database and an interface to that database written in PHP).

Collecting papers and links is optional: you can switch these features off and just use Postgenomic as a plain old blog aggregator.

Installation instructions are inside the repository - or you can read them here.

... there are more blogs in the index now, too.

I'm experimenting with new ways to rank blogs on the basis of the number of incoming links from other science blogs (and some secret mathemagic). We'll see how it pans out.

Other Postgenomic news:

• Egon has set up the excellent Chemical blogspace - nice.
  
• I've been going through some ideas with the API at Flags and Lollipops (a hot or not, popular posts widget and a Connotea mashup, so far).
• New code should be in the SourceForge svn repository by the end of the week.
  

The site hasn't been updated for about a week now. This is down to trouble with the aggregator not being able to pick out the metadata from papers correctly (I changed job recently and lost access to some journals that my previous workplace had a site licence for).

It'll take a wee while to get things up and running again. The aggregator itself is still running, so no blog posts will be lost.

It took three months, but Postgenomic now indexes and archives commentary on over a thousand papers (there's a caveat in that a few news items are still classed as papers). There are just over eight thousand blog posts in the database (and almost three thousand distinct tags - that's including blog categories and rel='tag's. So much for consolidation).

If you've submitted a blog in the past month or so, I apologise. Although Postgenomic is tiny by, say, Technorati standards it's getting quite difficult to maintain and update in its present form.

This is why it doesn't look like there's much happening to the site: when I've had time to implement any changes they've been to the backend. It's also why no new blogs have been added recently. I'll be updating the blog list in the next two or three weeks.

As part of that update there'll be a major change or two, like the introduction of fields outside of the life sciences. Physics will be first up, thanks to Jeff Hodge's excellent aggregator Mixed States (if you're even faintly interested in physics and haven't yet checked it out, you should).

Postgenomic already collects reviews - pieces of useful information concerning papers. We'd also like to collect conference "buzz": be it background information, ancedotes or even full blown conference reports.

The idea is to help people to get a feel for different conferences: were the talks good last year? Are there any hotly tipped speakers? Is it worth the registration fee?

You can mark conference related posts with a class='conference' attribute on the tag linking to the conference homepage, like so:

I just got back from <a href='www.biowidget.com' class='conference'>BioWidgetWorld '06</a>, where I saw a kickass presentation...

We'll show the first few sentences of each post together with a link to your blog.

(edit: the syntax has changed from rel="conference" to class="conference")

You can now set up "watchlists" on Postgenomic that will inform you via RSS if a particular paper or keyword pops up anywhere in the blogs we index.

This means that you can easily track, for example, papers that you've written (just enter your name in the "Any papers containing the text..." box). You can also track keywords to create a subject specific feed of life science blog posts - try typing "AIDS" or "H5N1" into the "Any posts containing the text..." box.

On a slightly more advanced note you can also track any papers in libraries on Connotea - just enter your Connotea username in the relevant box and Postgenomic will do the rest. This makes it very easy to track papers that you come across while browsing the site - all you need to do is click on the gray Connotea icon next to the paper title to add it to your library.

Postgenomic is in the process of moving servers, so access might be flaky over the weekend as the DNS changes get propagated. Everything should be back to normal by Tuesday at the latest.

We're moving over to a server owned by Seed Media Group (owners of Seed Magazine and ScienceBlogs) - hence the wee "hosted by Seed" button in the footer. They very generously offered to host the site for free after Postgenomic had server troubles at the beginning of this month - hopefully we've now got some room to grow.

I've been busy fixing bugs for the last wee while (with much help). The next fortnight, however, should see the introduction of a couple of new features. I'll discuss a few of them below.

The first new feature is "watchlists". The idea is that you can set up a list of papers - you can get Postgenomic to import your data from Connotea and some types of citation manager, which should make things easier - and then select which ones you want to keep an eye on. If a paper gets cited or reviewed by a blog, or its record on Postgenomic gets updated in some other fashion (it's reviewed by the Faculty of 1000, for example) then we'll inform you by email or RSS.

You could use it to keep track of discussions involving a paper you or your lab had a hand in.

The second new feature is clustering of news stories. Quite often the same popular story will have more than one link in the top 50 - ten people might blog it, but five will link to one newspaper reports and five to another. Clustering should pull everything together. We're using Stijn Van Dongen's excellent Markov Clustering Algorithm (MCL), which amongst other things is also used to cluster protein sequences into the protein families you see on Ensembl.

The third new feature is an API which will let you use Postgenomic content on your own blog and in your own scripts. Use the data any way you see fit (bearing in mind the rights of the original authors).

Things that are still in the planning stages include conference reports (how should they be collected?) and browser plugins / extensions (if you have any experience of writing these and fancy contributing to Postgenomic, please get in touch) so that users can read any feedback on papers as they come across them on the web.

Happy browsing.

When there are any major updates to the site, I'll post them here. I'll also post occasionally about tasks that need to be done, things that, in retrospect, should've been done differently, people who should be thanked - that sort of thing.