Our lab focuses on two major areas of Drosophila development:

Stochastic choices in development: One of the most exciting projects concerns the stochastic distribution of two types of unit eyes (ommatidia), which is reminiscent but evolutionary unrelated to the stochastic distribution of blue, red and green cones in the human retina. We discovered that the stochastic expression of a single transcription factor, Spineless, in a subclass of the R7 photoreceptors is responsible for this choice (Wernet et. al., 2006). This led us to generalize the notion that developmental stochastic choices are fairly wide spread in sensory systems. After studying in detail the genetic network that controls fate determination of each class of photoreceptors (Johnston et. al., 2011), we have recently provided a mechanistic understanding of how each allele of spineless makes an independent choice, while later communication between alleles across different chromosomes allows the coordination of expression within a given cell (Johnston & Desplan, 2014). We are now extending this work to look at the evolution of these stochastic choices in other insects. We have shown that butterflies and wasps have two R7 cells that make choices between blue and UV opsins, leading to four possible combinations that are entirely determined by spineless: when we disrupt spineless in butterflies or wasps with CRISPR, the two R7-like photoreceptor express a single opsin (Perry et. al., 2016).

Generation of neural diversity: the development of the optic lobes: The next logical step was to understand the development and function of the different neuronal types of the optic lobes. We first focused on the Medulla, the largest neuropil formed by 40,000 neurons. We defined the more than 70 cell types that compose the medulla, each surrounding one (or several) of 800 photoreceptor projections to form a precise ‘retinotopic’ map (Erclik et. al., 2017). We undertook understanding how these 70 cell types were produced by 800 neuroblasts and discovered a temporal series of 5 transcription factors expressed sequentially in neuroblasts that produce different neurons at each temporal window (Li et. al., 2013). This work provided a generalization of the temporal patterning model that was discovered in the Drosophila embryonic nerve cord. This opened the door to understanding temporal patterning in the vertebrate retina or cortex. We also discovered that the temporal factors not only determine neuronal fate but also death or survival: at a certain temporal window, NotchON neurons die by Hid-driven apoptosis while at other temporal windows, the NotchOFF progeny die through Reaper-induced apoptosis (Bertet et. al., 2014). Finally, we analyzed the function of specific neurons that participate in every one of the 800 columns. We discovered that two neurons mediate responses to bright edge motion, with one of them delayed as compared to the other as to allow coincidence detection of stimuli from two neighboring photoreceptors. Two other neurons do the same for dark edge motion. We were thus able to define the neurons that implement the famous Hassenstein-Reichardt Elementary Motion detector first described in the 50’s (Behnia et. al., 2014). Our current work is focused on analyzing the other optic lobe structures and on discovering novel mechanisms of neurogenesis involving different strategies that are comparable to the neurogenesis of different vertebrate brain structures.