RSS button

the latest news

Automation of fly behavioral assays

PDF version

Check out this nice article on recent advances in the automation of fly behavioral assays. The piece examines four automated assays: 1) from our lab, the LegTracker microscope, which allows us to track the position of all 6 of a fly’s legs at high frame rates, in real time, while it runs on a floating ball. 2) From David Anderson and Pietro Perona, an arena tracking assay used to examine social interactions of flies as they share a food resource. 3) From the lab of Carlos Ribeiro flyPAD, which uses capacitive sensors to monitor the feeding of individual flies with very high temporal resolution and sensitivity. In fact, even single gulps of a fly’s proboscis can be discerned. 4) From the lab of Barry Ganetsky, Fly Tester, a simple but brutally effective assay to measure flight capabilities – drop flies down a cylinder lined with sticky paper. Flies that regain flight control first will fly into the walls higher than weak flies that plummet to the bottom.

There are two aspects I particularly like about this piece. First, it considers explicitly the practicality of making each of these devices? How technically challenging are they to put together? What kinds of data collection does that technicality permit? How much does it cost to fabricate the instrument. Additionally, the article gave me some room to speculate on the origins of the current flourishing in fly automation. Specifically, I think the advent of cheap microelectronic sensors, CCDs, CMOSs, spectrally diverse LEDs, etc., was the main factor. The steady march of Moore’s law has meant that any post-processing of data from cameras can now happen in reasonable amounts of time, if not real time. And the maturation of machine learning has meant that even CS neophytes, like us, can throw together reasonable classifiers to parse behavioral data. All in all, a great time to be working in this field.

February 21, 2015 on 1:09 am    ~    No Comments

2013-2014 academic year in review

Well, our first year in the Northwest building has come and gone in a whirlwind. There hasn’t been a news update here in that time, and for that I only have the tepid excuse of being busy. From here on out, this space will hopefully be more interesting and lively.

As a recap of the activity over the last year, I’ll focus on who has come and gone in that time, what they worked on, and who makes up our current team. Starting with those who, to our great detriment, have already moved on.

Chelsea Jenney first joined the lab as an REU summer student in 2012. She worked with Julien and Sean to screen many different strains of flies, each of which was more or less genetically identical within a line but differed between lines. For each line, Chelsea measured the extent to which dozens to hundreds of individual flies differed in their locomotor behaviors, so that each line was characterized by how variable its behavior was within the line. She did this for more than 150 different strains, a massive undertaking resulting in a huge data set. All of these strains have been sequenced, so Julien was able to detect genetic polymorphisms (and associated genes) between the lines that predict higher or lower behavioral variability. Much of the last year, in which Chelsea was a research technician in the lab, was devoted to validating the effects of these variability genes. Chelsea wrapped up her technician stint in June, and we’ll miss her enthusiasm, good cheer and prodding to go get bahn mi for lunch.

Vivian Hemmelder joined the lab in January 2014 as a research intern during her gap year between undergraduate and graduate school. She’d spent the fall working in the Murthy Lab, also in CBS, on a project investigating how mice can smell a single odor against a background of many odors. In my lab she worked on an entirely different project – how do flies generate decisions in the absence of external stimuli driving their behavior? To study this, Vivian engineered a new rig in the lab, a combined Y-shaped maze and phototactic stimulator. As with our other rigs, this one is massively parallel to study many individual flies simultaneously. After hard work with microcontrollers and wiring up hundreds of LEDs, Vivian’s rig came to life and she started screening for circuits in the fly brain that affect this behavior. But, as these things go, time ran out right as the data was starting to flow in. Vivian headed home for the summer before she begins as a first year graduate student in the Program in Neuroscience this fall. Here’s hoping she brings her creativity and enthusiasm back to the lab in the future.

The lab has three permanent members and one associate at present.

Kyle Honegger was the first person to join the lab, approaching Ben while we were still at the Rowland with a brilliant project to determine how physiological variability in the encoding of odors within the olfactory system correlates with behavioral variability. Kyle did his PhD with Glenn Turner in the Cold Spring Harbor Watson School on the representation of odor information with the fly mushroom bodies, brain regions dedicated to olfactory associative conditioning. Kyle brings formidable optophysiology and behavioral talents to the lab.

Kyle’s research question falls in a new and essential category for us. For in addition asking what are the factors that regulate behavioral diversity and variability, Kyle is asking what specific factors within the brain make an individual’s behaviors different from his or her identical siblings’? What parts of the brain do we need to examine in order to make predictions about a specific animal’s behavioral tendencies?

Kyobi Skutt-Kakaria was the first graduate student to join the de Bivort lab! Kyobi comes from a molecular background and will be addressing a key question in our understanding of behavioral individuality – where in the molecular cascade from DNA to RNA to proteins to cellular physiology to information processing in circuits does behavioral individuality arise? Is it a developmental process? Can it be localized to specific genes or signaling pathways, and if so, what in what neurons do these processes act to determine behavior? Kyobi brings much needed expertise in molecular biology to the group and will be our resident deep sequencing guru. Congratulations to Kyobi for passing his MCO qualification exam this spring with flying colors!

Zach Werkhoven is the newest member of the lab, joining in late spring 2014. Zach’s rotation project was an engineering challenge: build a motor-controlled ball that an animal – not just flies, but any small terrestrial creature – can run on top of, while being held in a fixed position and orientation by the counter-rotation of the ball. This tool will allow us to non-invasively measure behavior of a wide (phylogenetically diverse) variety of insects – opening up the question of how behaviors evolve over long timescales. Does evolution routinely invent new behaviors, or does it re-purpose and re-tune a standard toolkit?

Julien Ayroles is an associate of the lab. He is a PI in his own right, as a Junior Fellow of the prestigious Harvard Society of Fellows. Julien is an affiliate of several lab groups at Harvard and beyond, and brings to each his own research program to investigate the genetic basis of phenotypic variability. His talents are diverse and deep, and he will probably only be around Harvard for a limited time before embarking to start his own lab.

We’re in luck to have two rotation students from the Program in Neuroscience in the lab through the end of the summer. Jess Kanwal is asking a fascinating question – do individual flies vary in simple behaviors that are executed by very simple circuits? Specifically, she is asking if flies vary in their jump response, which is triggered by brief flicks of darkness, a behavior mediated by a ~6 neuron deep circuit. If so, this presents a greatly simplified set of targets to examine to understand the origins of behavioral individuality. Michelle Frank is working with Kyle on a number of key experiments related to odor-response individuality. Does the mushroom body, which can pair positive and negative associations to odorants contribute to the behavioral diversity we see in response to odors? Alternatively, are odor-response differences facets of the innate odor-preference circuitry? Michelle is also helping to assemble and align our new and improved 2-photon optophysiology microscope.

Most recently Alex Isakov a physics student from L Mahadevan‘s group in Physics, Engineering, OEB, (etc.!) began work on a collaborative project to precisely measure and model how the walking leg coordination of flies response and (maybe) recovers from injuries. Alex put together a great team of summer students: Brian Sullivan and Akshitha Ramachandran to measure highish speed video of many different flies pre- and post- injury, and then score with meticulous care their leg-positions frame by frame.

Lastly, we have a collaboration going with Liz Kane’s group at the Rowland Institute at Harvard, namely Jeremy Todd, Dalhyung Kim and Jack Michaud to measure the locomotion of individual Drosophila larvae and ask if they show the same kinds of locomotor individuality as adults, and if so, do larval biases persist through metamorphosis into adulthood?

Stay tuned, with improved blog diligence there will be updates to these emerging stories in the near and medium term.

August 1, 2014 on 8:20 pm    ~    1 Comment

We’ve moved to OEB / CBS

I’m happy to say we’ve finished most of the move from the Rowland Institute to our new home in the Northwest building on the main campus!

The crates are now unpacked and we’ve resumed collecting data.

Our summer intern Sandra has been collecting Y-maze handedness data on flies reared under standard culture conditions, and flies reared in visually and spatially enriched environments. Our hypothesis: that environmental enrichment during development and early adulthood will increase the extent of inter-individual behavioral variability. What’s great about this hypothesis is that whether it, or its opposite, is true, we’ll have an interesting result. Both alternatives are plausible and have theoretical support:

  • Given circuit plasticity, and the possibility (in enriched environments) that each fly can find its own particular microenvironment, environmental complexity may increase behavioral variability.
  • But, environmental enrichment has been shown in numerous protocols, to increase the “robustness” of development, i.e. how closely to the standard blueprint each animal ends up. If behavioral variability is a symptom of developmental decanalization (a lack of robustness to environmental perturbation) then enrichment may diminish variability.
  • The results so far? Well, they’re complicated, but it looks like enrichment may increase the inter-individual variation in some traits, and have no effect whatsoever on others. It seems that there is substantial variation behavior-to-behavior in the extent to which environment can alter inter-individual variability. For some behaviors, the amount of inter-individual variability may be (paradoxically) hard-wired in the genome.

    The data is flowing in and it will be great to see this story emerge.

    The last view of the lab (in crates) at the Rowland Institute. Thanks Rowland, you were a wonderful incubator for my brain.

    July 24, 2013 on 12:51 am    ~    No Comments

    Next Page »