The file itself is MS Word doc. Those I save as native openoffice on my system. Now, an openoffice document is really just a bunch of mostly XML documents zipped together. If you do the following:

unzip  -l conference-delegates.odt

You get a listing that looks like this:

Archive:  conference-delegates.odt
 Length      Date    Time    Name
---------  ---------- -----   ----
 39     2010-09-01 18:16   mimetype
 71244  2010-09-01 18:16   content.xml
 94     2010-09-01 18:16   layout-cache
 15522  2010-09-01 18:16   styles.xml
 1241   2010-09-01 18:16   meta.xml
 24852  2010-09-01 18:16   Thumbnails/thumbnail.png
 0      2010-09-01 18:16   Configurations2/accelerator/current.xml
 0      2010-09-01 18:16   Configurations2/progressbar/
 0      2010-09-01 18:16   Configurations2/floater/
 0      2010-09-01 18:16   Configurations2/popupmenu/
 0      2010-09-01 18:16   Configurations2/menubar/
 0      2010-09-01 18:16   Configurations2/toolbar/
 0      2010-09-01 18:16   Configurations2/images/Bitmaps/
 0      2010-09-01 18:16   Configurations2/statusbar/
 8961   2010-09-01 18:16   settings.xml
 1988   2010-09-01 18:16   META-INF/manifest.xml
---------                     -------
 123941                     16 files

Wow. Which file contains the delegates’  emails in all that? Actually, content.xml contains the textual content of the openoffice.org document. You can open it with your favorite XML and see how it’s constructed (I like Firefox myself for browsing, and XML Copy Editor for more in-depth diagnosis). But for now, we would like to extract the emails. So we unzip content.xml only:

unzip conference-delegates.odt content.xml

This unzip command will only extract content.xml from the archive that is the .odt file.

When looking at the content.xml file, we see lines like this:

 <text:a xlink:type="simple" xlink:href="mailto:noone@usc.edu">

 <text:span text:style-name="Internet_20_link">
 <text:span text:style-name="T2">noone@usc.edu</text:span>
 </text:span>
 </text:a>
 </text:p>

Which means that “noone’s” (usernames have been changed to protect the innocent) email appears both as text and as hyperlink. It may or may not be that all the delegates’ emails are hyperlinked, so we may expect some duplications we need to get rid of.

To get the email addresses themselves, we use egrep. egrep uses the extended regular expression syntax in searching for emails. What is a good regex for email addresses? There is a good discussion of that at the regex-guru site. I use the rather simple form:

egrep -o -i '\b[A-Z0-9._%+-]+@[A-Z0-9.-]+\.[A-Z]{2,4}\b' content.xml

Explanation: the  -o qualifier prints only the word matching the regex.  -i means a  case-insensitive match. egrep, the extended version of grep, that can handle regexs with things like {m,n} repeats. However, the result of our little exercise would still have duplicate emails, because of the hyperlinking tags. Here is how to get rid of the duplicates:

egrep -o -i '\b[A-Z0-9._%+-]+@[A-Z0-9.-]+\.[A-Z]{2,4}\b' content.xml | sort |  uniq

sort sorts the output alphabetically, preparing it for uniq to get rid of duplicates.

One last touch-up: we really don’t need to physically extract the content.xml file. “unzip -c” extracts files to stdout. Therefore, we can get the email addresses without cluttering our disk:

unzip -c conference-delegates.odt content.xml | \
egrep -o -i '\b[A-Z0-9._%+-]+@[A-Z0-9.-]+\.[A-Z]{2,4}\b' | sort |\
uniq > email-these.txt

Voila! email-these.txt now contains the emails of the conference delegates.

One last word: it may have been easier just to save the MS-Word doc file as text using the File -> Save as…” option in openoffice.org. Supposed we saved the file as conference-delegates.txt. We wouldn’t have to muck about with all the XML, and remove the email address duplicates due to hyperlinking. We could have just done:

egrep -o -i '\b[A-Z0-9._%+-]+@[A-Z0-9.-]+\.[A-Z]{2,4}\b' \
conference-delegates.txt > email-these.txt

But where’s the fun in that?

Happy spamming!

Not Predator MX.

M. xanthus is a highly cooperative bacterium, as we have already seen: when starved, most cells “commit suicide” while a few form spores, to survive the lean times. But M. xanthus also cooperates when times are good and food plentiful: M. xanthus cells form swarms, which envelope their prey and increase the concentration of digestive enzymes they secrete to the environment.

M. xanthus also ripple together to better ingest the nutrients released after digesting their prey. To do so, they use a type of pilus (motility organ, like a bacterial tentacle) called type IV secretion pili. But amazingly enough, in mutant M. xanthus that are unable to make these pili, a new mechanism for cooperative swarming evolved.  In 2003, Velicer & Yu from the University of Tuebingen have used a group of mutant M. xanthus, which are lacking the type IV pili gene  to show that cooperative swarming can evolve using an alternative mechanism. The mutant bacteria do that by forming a physical net of sugars and proteins connecting them — and their rippling motion — together. Watching behavior evolve: how cool is that?

Finally, a movie of M. xanthus swarming & rippling:

Velicer, G., & Yu, Y. (2003). Evolution of novel cooperative swarming in the bacterium Myxococcus xanthus Nature, 425 (6953), 75-78 DOI: 10.1038/nature01908

Berleman, J., Chumley, T., Cheung, P., & Kirby, J. (2006). Rippling Is a Predatory Behavior in Myxococcus xanthus Journal of Bacteriology, 188 (16), 5888-5895 DOI: 10.1128/JB.00559-06

Life cycle of M. xanthus. Credit: Carla canales / citizendium.org

This post deals with another aspect of bacterial cooperation: how does it evolve? Why cooperate in the first place at all? Every time an individual cooperates, short term gains are sacrificed for long-term ones, but those long-term ones are contingent upon all or most cooperating individuals doing their bit. Think about standing in line to the bus. If everyone cooperates, we get on the bus faster, but some of us may be forced to stand. On the other hand, shoving your way to the beginning of the line will assure you a good seat, albeit at the expense of glares from your fellow-passengers, and maybe a few altercations along the way.  In evolutionary terms, selfishness seems like a sounder strategy than cooperating.  After all, if you manage to gain a better position for yourself in life’s pecking order, you pass those genes that enable that to your progeny, and further down the line. Why cooperate or act selflessly in the first place? Why let someone else share the gene pool with you when you can have it all to yourself?

Unless that “someone else” shared genes with you: that is, they were related in some way. Suddenly, cooperation seems to have evolutionary benefits: you are preserving and passing on some of the same genes.  Protecting kin is the most often-used explanation for how cooperation evolved in the first place: kin selection, meaning, favoring cooperation those individuals with which you share a larger number of genes over those who do not. Evolutionary biologists use the Hamilton’s law as a guideline:  the higher the benefit of the cooperation, the lower the cost, and the closer the relatedness of the individuals cooperating, the more likely it is that there will be cooperation. Putting it into an equation, cooperation will evolve if the following condition is met:

rb - c > 0

Where r is the relatedness (on a scale of 0 to 1 where 0: no genetic relation, 1:self), b is the benefit of cooperation and c is the cost. This rule, formulated by William Donald Hamilton is a centerpiece of evolutionary biology. Imagine going on a day’s hunt  where the quarry is a large animal that can feed one hunter for 35 days, but requires at least five hunters to take it down.  Now there are also smaller animals around, that can be hunted by one individual, and they supply enough food for one hunter for two days. Is it beneficial to hunt  alone or together? Let’s figure the benefits and costs. For hunting the large animal, the one that requires at least five hunters, the benefit is a week of food each (b= 35/5 = 7) while hunting for one day (c=1). If the individuals are cousins sharing an average of 20% of the genetic material  then:

0.2×7 – 1 = 0.4 is the benefit score

If they are siblings, sharing 50% of the genetic material, then:

0.5×7 – 1 = 2.5 the benefit score is even higher

But what about individual hunting? The benefit of the smaller quarry is is 2 days worth of food, and one day of hunting, and you do it alone. So r=1 (yourself), b=2 and c=1 giving us:

1×2-1 = 1

In this hypothetical model, a group of siblings will cooperate to hunt big game, while cousins would probably hunt smaller game alone. If you want to dig deeper into how Hamilton’s rule was derived, and further implications of the rule, I recommend this post.

Any mechanism in evolution is examined through the lens of fitness. Fitness is the relative ability to produce and support viable progeny. So if cooperation increases fitness, we can use the following simple graph to explain the difference between a cooperating and a non-cooperating  individuals in a cooperating population using Hamilton’s rule:

Figure 1: Hamilton's rule prediction: the fitness of cooperators (blue) and non-cooperators (red) increases as the number of cooperators among social neighbors (x-axis) increases. The slope of both lines is the benefit (b).

The benefit, b, is the slope of these two lines. The difference is c. Note that for any given frequency of cooperation in the population, the non-cooperating individuals (red line) have a higher fitness than the cooperating ones (blue line). It seems that it “pays off” to be a self-server no matter the social environment you are in, even though you still benefit from being in a cooperating community. Yeah, we all know the type.

But what happens when the difference between cooperating and not cooperating depends on the percentage of cooperators in the population? Not too hard to imagine: if most of the population is playing nicely together and benefiting from it, then this might change the attitude of the selfish individuals more readily then if only a small fraction of the population is cooperating. But as it stands, Hamilton’s rule does not provide for this type of model. However, the following modification of Hamilton’s rule does:

r ⋅ b – c + m ⋅ d > 0

Relatedness, r, is now not a scalar (a single number), but a vector (an ordered set of values)  r = {r₁, r₂, …} describing relatedness under different conditions. Ditto the benefit vector, b. b has the coefficients of the equation describing the fitness of non-cooperators as a function of how many neighboring cooperators there are in the population (red lines). In a linear setting (Figure 1), r = {r₁} b={b₁} and m⋅d = 0, collapsing the expanded equation into the classical Hamilton’s rule.  We won’t get into m and d in this post, they are important though, and you should read the paper to understand how they play a role

Expanding them from scalars to vectors enables a richer and more flexible description of Hamilton’s rule, allowing to describe non-linear relationships like this:

Figure 2: Note two things. First, the relationship between fitness and the fraction of cooperators in the population is not linear. Second, the difference in fitness between cooperators and non-cooperators decreases as the fraction of cooperators in the population goes up. These two phenomena cannot be described by the classic Hamilton's rule equation. They can be described using the modified rule.

This modification of Hamilton’s rule was developed by Jeff Smith and colleagues, at the department of Biology at Indiana University. Armed with the new equation, Smith and his colleagues decided to see how well it can be applied. They decided to look at Myxococcus xanthus. M. xanthus bacteria behave normally as long as food is abundant: they swim around and proliferate by cell-division as bacteria do. But when starved, they aggregate, and some cells form resistant spores, while the others die. Some cheating strains sporulate  well when in cooperating populations, but do poorly on their own. The scientists mixed a cooperator strain with a cheating strain at different frequencies, starved them, and measured the fraction of each strain in the population of surviving spores. They found the following: first, the fitness effect was non-linear; in fact, it was almost exponential. Second, cooperators were more fit than cheaters at low cooperator frequencies, but cheaters fared better at high cooperator frequencies. So it pays to freeload when most people around you behave nicely. In the case of M. xanthus, the added value to the population is quite high. In fact, the scientists found that cooperation in M. xanthus is very robust and resistant  to cheating:  cheaters were viable (i.e. had a positive fitness)  only with groups that had more than 70% cooperators. So it is only when cheaters have a large cooperating population to buffer their nasty habits that a they can thrive.

Figure 3: Relative fitness of cooperators (blue) and cheaters (red) in a populations with different relative frequencies of cooperators. Note that the fitness scale is logarithmic: the fitness increase is very much non-linear, as in Figure 2.

Moral of this story: if you got to cheat, make sure there are a lot of nice people around. Otherwise it won’t work out very well.  In evolutionary terms, the  trait for cooperation and kin selection has evolved to become strongly entrenched, so much that cheaters can only survive if cushioned by a high frequency of cooperators. Favoring your own and acting selflessly towards them seems to be the way to go, in the case of M. xanthus.

Why am I harping on this again?  Because the term “low homology” managed to sneak itself, of all places, into the title of a paper published in Bioinformatics.  Ouch. Bioinformaticians should know better.

Just for kicks, I decided to look at how many papers were published this year (January 1 through today)  using the misuse of terms in their title or abstract. Here are the results:

• “high homology”  134
• “low homology”: 13 (well, that’s low)
• “highly homologous”: 140
• “distant homologs”: 7
• “close homologs”: 7
• “percent homology”: 1

I could not find others such as “weak homologs”, “strong homologs”. Small mercies. Well, there is some work to do still in removing bad habits.

Update: the video embedding has been disabled. You can still watch it on YouTube.

Some headlines just write themselves…

It has been known for some time that an approaching large herbivore causes aphids to abandon ship …err plant. Makes sense since, after all, there’s not much of a point in staying on the particular bit of shrubbery that will be consumed, lock, stalk and barrel by a ravenous forager. However, it was not exactly clear what in the herbivore causes the aphids to drop. Well, it is not the shaking of the twigs, as rustling the plant did not cause a substantial number of the aphids to drop. Rather, it’s the breath. The researchers had a human, a sheep and goat all breath on an aphid-infested plant, with equal results: the aphids dropped from the plant en-masse. But what in the breath causes aphids to do that? Well, it is not the CO₂ nor the air movement itself. Rather, the heat and the humidity of the breathing, as tested by Moshe Gish and his colleagues at the University of Haifa.

This is a great example of adaptation: after all, bush movement may be due to many different factors, including uninterested rodents and carnivores. Also, air movement can be simply caused by wind, including hot or humid air.  But someone breathing directly on you, hot and damp can only mean one thing to an aphid: abandon plant or be goat dinner!

A few interesting facts about sloths (edited from Wikipedia):

• Look at your forearm. Your hair grows towards your hand. In most mammals hair grows towards their extremities. In sloths, hair grows in the opposite direction.
• Sloth metabolic rate is very slow, their body temperature is as low as 30C, and they move very slowly and only when necessary. Which makes them perfectly suited for government jobs
• Because of their low metabolic rate, they only go to the bathroom once a week.
• Almost all mammals have seven cervical vertebrae (neck vertebrae). This has nothing to do with neck length: both a giraffe and the narrator at the Rocky Horror Picture Show have seven neck vertebrae. Two-toed sloths have only six cervical vertebrae, and three-toed sloths have nine cervical vertebrae

Meet the sloths from Amphibian Avenger on Vimeo.