Hello! Recently, I’ve been tweeting writer’s advice from @jbwallingford using the hashtag: #DevBiolWriteClub. I’m psyched that The Node is now letting me add a little depth to this venture. In this first post, I’ll start by managing some expectations.
If you’ve followed #DevBiolWriteClub on Twitter, you might recall that one of my earliest tweets said: “For now, I will focus on how to be a better writer. This is different from how to write better.” Personally, I like that sentence. I think it’s clever, and it conveys some crucial information. That said, one could also make the argument that it sucks. After all, it’s kind of confusing, and by forcing the reader to think hard about just a few words, it risks failing to convey any information at all.
Regardless, the details of well-written and poorly written passages is NOT the point of #DevBiolWriteClub. Rather, I want to use this forum to address one of the most common and intractable misconceptions in writing, especially among busy scientists. The issue is this: If you are serious about better writing, DO NOT start by thinking harder about sentence structure and grammar. Instead, start right now by focusing on your practice of being a writer.
Writing is like a sport. You only get good at it if you practice, with intent, every day. When I’m not doing science, I’m a rock climber, and one of my favorite coaches is Steve Bechtel (Climbstrong.com.), and he once lamented that he always wanted to write an article titled “500 weeks to stronger fingers,” but that no one would read it. I fear the same for these blog posts, because he’s exactly right. No one likes to hear that there are no shortcuts.
The difference, of course, is that most scientists relate to sports (or art, or baking) as a hobby, as this other thing we do sometimes, and we hope to get better at. Ultimately, though, if we don’t get better, it’s not that big a deal. The problem is that many scientists take the same view of writing. But not improving as a writer is a big deal in science. It’s a very big deal, actually. If you are a scientist and you want to succeed, you mustbecome a writer. And the only way to do that is to practice, day in and day out. For years.
So, that’s the bad news. The good news is that the process for becoming a writer is pretty simple, and you can start today. There are only five rules:
Do the work.
Do the work.
Revise and edit. Again, and again, and again.
Read with intent.
You can’t do it alone.
Now, it’s essential that you understand these rules, especially #1 and #2:
Of course, writers need to write, but what I mean by “do the work” is broader. I want you to create a new habit of mind. Take time out of every workday to practice the craft of being a writer by following any one of the five rules. What you do in each session is less important than doing something each workday.
Maybe you will actually write new words in a session, but if you don’t, that’s OK! Some days are stacked with experiments, so maybe you can only find a few minutes to revise something you’re working on (#3). Or maybe all you can manage is to spell-check something you wrote yesterday (also #3). Or maybe you just read (#4). Or maybe today’s the day you have the courage to show what you’ve written to a friend and get their feedback (#5). All of these examples can fall under Rules #1 and #2. But here’s the thing: You need to approach every session intent. Set aside time to do it; try hard; and when you’re done, reflect on your performance. This is the work. Let’s get to it.
In the coming posts, I’ll write more about each of these rules and I’ll provide advice on how to follow them.
In the latest episode of Genetics Unzipped, Kat Arney sits down with leading evolutionary geneticist Professor Sarah Tishkoff from the University of Pennsylvania to talk about her work mapping the genetic diversity of African populations. Sarah talks about the challenges of carrying out ethical fieldwork in Africa and explains why its so important to us more genetically diverse data to inform the development of precision medicine.
We also hear from Garrett Hellenthal and Lucy van Dorp from the UCL Genetics Institute, who are unearthing hidden histories and cultural complexities hidden within DNA. From the mighty Kuba Kingdom to the curiously exclusionary Ethiopian Ari people, the genomes of these populations tell rich, detailed stories about people and places.
If you enjoy the show, please do rate and review on Apple podcasts and help to spread the word on social media. And you can always send feedback and suggestions for future episodes and guests to podcast@geneticsunzipped.com Follow us on Twitter – @geneticsunzip
Flight feathers are amazing evolutionary innovations that allowed birds to conquer the sky. A study led by Matthew Towers (University of Sheffield, UK) and Marian Ros (University of Cantabria, Spain) and published in the journal Developmentnow reveals that flight feather identity is established thanks to Sonic hedgehog – a signalling molecule well-known for giving the digits of the limb their different identities (so that your thumb is different from your pinky, for example). These findings suggest the pre-existing digit identity mechanism was co-opted during the evolution of flight feathers, allowing birds take to the air.
Feathers and the flight they support have long fascinated humans. In the bird embryo, feathers begin as buds—thickenings of the epidermis—that then develop into follicles, from which the keratin-based feathers are produced. Not all feathers are equal, however—compare for instance the downy feathers on the breast of a robin with the flight feathers of its wing. Classical embryological experiments in the 1950s, which involved grafting one part of the embryo on to another, suggested that feather identity (e.g. the choice to become a down versus a flight feather) is established at the earliest stages of development, even before the feather buds form. But in the seventy-odd years since then, we still don’t know much about which signals regulate feather identity.
The new study, carried out with Lara Busby as first author, reveals that flight feather identity is specified by Sonic hedgehog (Shh), a famous signalling molecule known to be involved in the development of limb digits, including human fingers. (And yes, Shh is named after the computer game character, but that’s another story.) Using chicken embryos, the scientists found that Shh is required in the earliest stages of wing development for the mature birds to develop flight feathers. They also defined a set of genes that are likely to be involved in this process. Importantly, they discovered that Shh works in a defined temporal sequence to specify the different flight feather identities, mirroring how it specifies the different digit identities. This similarity suggests that the digit identity network was co-opted for flight feather development during evolution.
Sonic hedgehog is normally produced at the posterior margin of the embryonic chicken wing bud. Grafts of Sonic hedgehog-producing cells were made to the anterior side of a wing bud of another chicken embryo. This operation duplicates the tissues of the mature wing, including the black-pigmented feather buds as shown in the image. The flight feathers buds are the ones protruding from the left and right margins of the wing. Duplicated tissues are on the anterior side of the image (left), and normal tissues are on the posterior side (right)
Dr. Towers said: “Flight feathers are one of the most important evolutionary adaptations that allowed birds to take to the air. Our unexpected findings, showing that the digits and flight feathers share remarkably similar developmental programmes, provide important insights into how the bird wing evolved to permit flight.”
The researchers hope to extend this work by trying to understand how the early exposure of embryonic chick wing bud cells to Shh is ‘memorised’ to allow flight feather formation at a much later stage of development.
Contact: Dr. Matthew Towers, Reader of Developmental Biology and Wellcome Senior Fellow in Basic Biomedical Science at the University of Sheffield, UK.
In a previous blog I explained how animated plots can be made to illustrate the dynamics of data. Animated plots go nicely together with the movies from which the data was extracted. Here, I explain how to display a movie and plot side-by-side, starting from a stack of images and using only open source software. I will use ImageJ/FIJI for image analysis and for creating the movie. Then, R/Rstudio is used for data processing, generating the animated plots and for adding the plot to the movie. A basic understanding of the software is assumed and necessary to complete the tutorial. The steps that are involved are 1) creating a movie, 2) quantification of signal over time, 3) labeling the movie, 4) creating an animated plot and 5) combining the plot with the movie.
The data
First, let’s discuss where the data came from. I acquired the data with TIRF microscopy, which is a technique that detects fluorescence at the basal plasma membrane of cells. TIRF imaging of cells expressing a GFP-tagged Protein Kinase C (GFP-PKC) was used to study the dynamic association of PKC with the plasma membrane, after stimulating the cells. The fluorescence intensity follows an oscillatory pattern that was triggered by adding UTP and that reflects calcium oscillations (Oancea and Meyer, 1998). The raw data, processed data and an R-scripts to generate the movie is available in a file archive at Zenodo.org, doi: 10.5281/zenodo.3785592
Step 1: Creating a movie
The data is astack of images named ‘EGFP-PKCbetaII_100uMUTP_16bit.tif’ which can be downloaded from the archive at Zenodo, or directly through this link. The TIF file can be opened with FIJI or ImageJ. To play the image stack as a movie press backslash [\] and again to stop it.
The first processing step is to convert the 16-bit stack to 8-bit (the line below refers to a ‘menu path’ like this: menubar > submenu > command):
Image > Type > 8-bit
A false color or lookup table (LUT) can be added to improve the visualization of the intensity changes:
Image > Lookup Tables > Fire
The ‘Fire’ LUT works well for this data, but other LUTs can be selected as well. To get the best result, several different LUTs should be tried. Here, I’ll use one of my favorites: the Parrot LUT, reported in this blog. The stack can be saved as gif to generate a movie:
The analysis should be done on the original data (not on converted 8-bit data). After opening the original TIF stack, regions of interest (ROIs) can be drawn to select individual cells. In this example I have drawn three ROIs in the center of the image. The ROIs are available here as a file called ‘RoiSet.zip’’ and can be added to the image stack by dropping the zip file on the ImageJ/FIJI menu bar. The ROIs are now available in the ROI manager that can be activated by selecting it under ‘Window’:
Menu > ROI Manager
Select all three ROIs in the ROI manager. Next, quantify the intensity of all stacks by selecting on the ROI manager window:
More>> Multi Measure
Make sure to activate the checkboxes ‘Measure all 100 slices’ and deselect ‘One row per slice’ and hit OK. A screenshot of the window and the Results window is shown below:
To indicate the cells that are used for the quantification, I have added labels to the GIF that was created in step 1. This can be done in ImageJ/FIJI:
Image > Stacks > Label…
An example of the settings and the result is shown in the screenshots below. Note that I have indicated that the label is only shown in the first 5 frames by setting ‘Range: 1-5’:
After adding the three labels, the image is saved again as a GIF. The result is available as ‘EGFP-PKC-labeled.gif’. I took note that the image has a height of 529 pixels, a number that we need later on.
Step 4: Animated Plots
For the next steps we switch to R/Rstudio. The complete script that I used is available as ‘Movie-with-plot.R’. Below, I will explain every step of the code. Running the code from the command line in Rstudio requires a bit of knowledge of R and it requires several packages that need to be installed and loaded (The ‘>’ sign indicates the prompt of the command line):
Now we are ready to load the data that was generated in ImageJ/FIJI:
>df_results <- read.csv("Results.csv")
The CSV file has a column named ‘Label’ that we need to split in three different columns based on a colon as a delimiter. The relevant column that we need later on is ‘Sample’:
I rename the column ‘Slice’ to ‘Frame’, since I think that is more appropriate.
>df_tidy <- df_tidy %>% rename(Frame = Slice)
To compare intensities, it makes sense to perform a normalization (also explained in another blog). Here, I normalize the intensity data by dividing the data (per Sample) by the average of the first 5 datapoints, reflecting a baseline. The data is stored in a new column, ‘Normalized Intensity’:
To synchronize the movie with the animated plot, it is essential to keep the number of frames (timepoints) the same. This number is determined from the data and we will use it later on:
>nframes <- length(unique(df_tidy$Frame))
Tweaking the layout
The appearance of the default plot can be fine-tuned in many ways. Below are a few of my favorite adjustments to the standard layout. The result of each step can be inspected in Rstudio by calling the object (note that it will take some time to render the animation):
When both the animated plot and the movie are ready, they can be combined. We will adjust the dimensions of the animated plot (size and number of frames) to that of the movie. The number of frames (nframes) has already been determined. The size of the movie can be assessed in ImageJ or FIJI. In this example, we combine the movie and plot horizontally, so we need to know the height of the movie (which is 529 pixels). With these parameters we can correctly render the animated plot and assign the result to object panel_b:
This object can be displayed by entering ‘combined_gif’ in the command line:
>combined_gif
Subsequently, we add all the other frames to the new object:
>for (i in 2:nframes) {
combined_panel <- image_append(c(panel_a[i], panel_b[i]))
combined_gif <- c(combined_gif, combined_panel)
}
Finally, we can save the object combined_gif as a GIF:
>image_write_gif(combined_gif, 'Combined.gif')
The resulting GIF is shown below:
Final Words
The combination of a movie and plot is an attractive and informative way to visualize data from timelapse imaging. This walk-through discusses the basics and lots of additional tweaking and annotations can be done. For instance, labeling of the images can be done in R instead of ImageJ, but I haven’t really figured out how to do that. A potential improvement would be to generate a script for the entire workflow (including the labeling), since it will make it simpler to generate movies in the long run.
I hope that this tutorial is useful and I’m curious how it works for you. I’d appreciate any feedback or suggestions for improvements and I’m looking forward to watch your movies and their plot!
We are excited to announce that the UK South West Zebrafish Meeting 2020 will be hosted by the University of Exeter as a virtual meeting to take place on Friday 11th September 2020.
The aim of the meeting is to bring together zebrafish researchers and technical staff from research institutions from the South West of the UK to share exciting cutting-edge research, knowledge, and zebrafish related expertise. We hope the meeting will forge new collaborations and networks across the South West and we hope you will be able to ‘join’ us! Registration and abstract submission are now open!
Due to the uncertain times regarding COVID-19 and possible long-term travel restrictions, we have decided to host our meeting online. We are in the process of working out the online format and logistics – so that we can host the best meeting possible – and we will inform participants of further details in due course.
We are also pleased to announce that we have 3 confirmed keynote speakers: Prof. Catherina Becker (University of Edinburgh), Dr. Isaac Bianco, (University College London), & Prof. Charles Tyler (University of Exeter).
Please pass the exciting news on to anyone you think may be interested and feel free to tweet about the event using #SWZM20 and follow us on Twitter @swzm20!
Looking forward to hearing from you all,
Steffen, Lucy, Holly, Chengting, Josh, Yosuke, and Michael
The Edmond and Lily Safra Center for Brain Sciences (ELSC) builds upon Hebrew University‘s record of excellence and innovation in its multidisciplinary approach to brain sciences.
ELSC invites applications for postdoctoral fellows in the following fields: theoretical and computational neuroscience, systems neuroscience, molecular and cellular mechanisms, cognitive neuroscience, and neuronal circuits. Postdoctoral fellows receive a competitive stipend for a period of up to two years.
Established ties and frequent collaborations with world renowned labs
Opportunities to audit advanced courses
Rich student and postdoctoral environments
Postdoctoral support staff
Eligibility:
The candidate must be (or have been) a student in an accredited institution of higher education and whose PhD training and post-doctoral projects are in the field of Brain Sciences.
The candidate’s doctoral degree has been submitted in the current year of applying or will be approved by the following year.
Candidates Commitments:
A recipient of an ELSC Fellowship must commence his/her post doctoral training no later than 5 years after completion of the PhD.
A recipient of an ELSC Fellowship must provide written approval from the authority of PhD students in his/her institute, confirming that his/her PhD has been submitted before they begin their post-doctoral training. If PhD was not yet awarded, the candidate must provide approval of a PhD during the first academic year of the post doctoral studies
A letter from the host is mandatory in order to commence the post doctoral studies
A recipient of an ELSC Fellowship must begin the postdoc training within 6 months after receiving the acceptance letter
Terms of Fellowship:
The fellowship can be extended up to 2 years, given availability of funds and the scientific achievements of the candidate. ELSC is not committed to prolong the fellowship in advance.
Preference will be given to students who completed their PhD abroad
We are looking for a motivated PhD student to join our lab recently established at the PMC in Utrecht. The position is funded for 4 years with a full-time employment. The project will deal with developing new (3D) culture systems and to use them to model cancer and study the effect of mutations, as well as organ development. We use state of the art tools, like CRISPR-Cas systems, single cell sequencing, mouse genetics and human organoids derived from both iPS cells and tissue stem cells.
The Prinses Maxima Centrum is a research institute dedicated to basic and applied research for different aspects related to paediatric cancer (https://research.prinsesmaximacentrum.nl/en/).
Ideal candidates have recently completed their master education, have a background in molecular and cellular biology, and preferably with a knowledge of bioinformatic tools, but this is not a must.
The candidate should be talented and highly motivated, and willing to work as part of a team. The group and the institute are international and the work language is English.
To have more detailed information about the project, the position and the lab please send a full CV and a brief cover letter to:
Welcome to our monthly trawl for developmental biology (and related) preprints.
With COVID-19 having shuttered labs around the world, it might be a surprise that April saw the highest ever monthly number of preprints deposited to bioRxiv: 3413, according to Rxivist. Though maybe not much of a surprise, with researchers at home and writing rather than pipetting.
In that 3413 we found the following couple of hundred developmental biology and adjacent picks (plus a handful from arXiv) – let us know if we missed anything. Use these links to get to the section you want:
Human cortical expansion involves diversification and specialization of supragranular intratelencephalic-projecting neurons
Jim Berg, Staci A. Sorensen, Jonathan T. Ting, Jeremy A. Miller, Thomas Chartrand, Anatoly Buchin, Trygve E. Bakken, Agata Budzillo, Nick Dee, Song-Lin Ding, Nathan W. Gouwens, Rebecca D. Hodge, Brian Kalmbach, Changkyu Lee, Brian R. Lee, Lauren Alfiler, Katherine Baker, Eliza Barkan, Allison Beller, Kyla Berry, Darren Bertagnolli, Kris Bickley, Jasmine Bomben, Thomas Braun, Krissy Brouner, Tamara Casper, Peter Chong, Kirsten Crichton, Rachel Dalley, Rebecca de Frates, Tsega Desta, Samuel Dingman Lee, Florence D’Orazi, Nadezhda Dotson, Tom Egdorf, Rachel Enstrom, Colin Farrell, David Feng, Olivia Fong, Szabina Furdan, Anna A. Galakhova, Clare Gamlin, Amanda Gary, Alexandra Glandon, Jeff Goldy, Melissa Gorham, Natalia A. Goriounova, Sergey Gratiy, Lucas Graybuck, Hong Gu, Kristen Hadley, Nathan Hansen, Tim S. Heistek, Alex M. Henry, Djai B. Heyer, DiJon Hill, Chris Hill, Madie Hupp, Tim Jarsky, Sara Kebede, Lisa Keene, Lisa Kim, Mean-Hwan Kim, Matthew Kroll, Caitlin Latimer, Boaz P. Levi, Katherine E. Link, Matthew Mallory, Rusty Mann, Desiree Marshall, Michelle Maxwell, Medea McGraw, Delissa McMillen, Erica Melief, Eline J. Mertens, Leona Mezei, Norbert Mihut, Stephanie Mok, Gabor Molnar, Alice Mukora, Lindsay Ng, Kiet Ngo, Philip R. Nicovich, Julie Nyhus, Gaspar Olah, Aaron Oldre, Victoria Omstead, Attila Ozsvar, Daniel Park, Hanchuan Peng, Trangthanh Pham, Christina A. Pom, Lydia Potekhina, Ramkumar Rajanbabu, Shea Ransford, David Reid, Christine Rimorin, Augustin Ruiz, David Sandman, Josef Sulc, Susan M. Sunkin, Aaron Szafer, Viktor Szemenyei, Elliot R. Thomsen, Michael Tieu, Amy Torkelson, Jessica Trinh, Herman Tung, Wayne Wakeman, Katelyn Ward, René Wilbers, Grace Williams, Zizhen Yao, Jae-Geun Yoon, Costas Anastassiou, Anton Arkhipov, Pal Barzo, Amy Bernard, Charles Cobbs, Philip C. de Witt Hamer, Richard G. Ellenbogen, Luke Esposito, Manuel Ferreira, Ryder P. Gwinn, Michael J. Hawrylycz, Patrick R. Hof, Sander Idema, Allan R. Jones, C.Dirk Keene, Andrew L. Ko, Gabe J. Murphy, Lydia Ng, Jeffrey G. Ojemann, Anoop P. Patel, John W. Phillips, Daniel L. Silbergeld, Kimberly Smith, Bosiljka Tasic, Rafael Yuste, Idan Segev, Christiaan P.J. de Kock, Huibert D. Mansvelder, Gabor Tamas, Hongkui Zeng, Christof Koch, Ed S. Lein
Connectomes across development reveal principles of brain maturation in C. elegans
Daniel Witvliet, Ben Mulcahy, James K. Mitchell, Yaron Meirovitch, Daniel K. Berger, Yuelong Wu, Yufang Liu, Wan Xian Koh, Rajeev Parvathala, Douglas Holmyard, Richard L. Schalek, Nir Shavit, Andrew D. Chisholm, Jeff W. Lichtman, Aravinthan D.T. Samuel, Mei Zhen
Spatial Transcriptional Mapping of the Human Nephrogenic Program
Nils O. Lindström, Rachel Sealfon, Xi Chen, Riana Parvez, Andrew Ransick, Guilherme De Sena Brandine, Jinjin Guo, Bill Hill, Tracy Tran, Albert D. Kim, Jian Zhou, Alicja Tadych, Aaron Watters, Aaron Wong, Elizabeth Lovero, Brendan H. Grubbs, Matthew E. Thornton, Jill A. McMahon, Andrew D. Smith, Seth W. Ruffins, Chris Armit, Olga G. Troyanskaya, Andrew P. McMahon
Atypical neurogenesis in induced pluripotent stem cell (iPSC) from autistic individuals
Dwaipayan Adhya, Vivek Swarup, Roland Nagy, Lucia Dutan, Carole Shum, Eva P. Valencia-Alarcón, Kamila Maria Jozwik, Maria Andreina Mendez, Jamie Horder, Eva Loth, Paulina Nowosiad, Irene Lee, David Skuse, Frances A. Flinter, Declan Murphy, Grainne McAlonan, Daniel H. Geschwind, Jack Price, Jason Carroll, Deepak P. Srivastava, Simon Baron-Cohen
The enhancement of activity rescues the establishment of Mecp2 null neuronal phenotypes
Linda Scaramuzza, Giuseppina De Rocco, Genni Desiato, Clementina Cobolli Gigli, Martina Chiacchiaretta, Filippo Mirabella, Davide Pozzi, Marco De Simone, Paola Conforti, Massimiliano Pagani, Fabio Benfenati, Fabrizia Cesca, Francesco Bedogni, Nicoletta Landsberger
Differentiated neural cells in Romero-Morales, et al.
Pod indehiscence in common bean is associated to the fine regulation of PvMYB26 and a non-functional abscission layer
Valerio Di Vittori, Elena Bitocchi, Monica Rodriguez, Saleh Alseekh, Elisa Bellucci, Laura Nanni, Tania Gioia, Stefania Marzario, Giuseppina Logozzo, Marzia Rossato, Concetta De Quattro, Maria L. Murgia, Juan José Ferreira, Ana Campa, Chunming Xu, Fabio Fiorani, Arun Sampathkumar, Anja Fröhlich, Giovanna Attene, Massimo Delledonne, Björn Usadel, Alisdair R. Fernie, Domenico Rau, Roberto Papa
Nuclear myosin VI regulates the spatial organization of mammalian transcription initiation
Yukti Hari-Gupta, Natalia Fili, Ália dos Santos, Alexander W. Cook, Rosemarie E. Gough, Hannah C. W. Reed, Lin Wang, Jesse Aaron, Tomas Venit, Eric Wait, Andreas Grosse-Berkenbusch, J. Christof M. Gebhardt, Piergiorgio Percipalle, Teng-Leong Chew, Marisa Martin-Fernandez, Christopher P. Toseland
Drosophila Sex Peptide Controls the Assembly of Lipid Microcarriers in Seminal Fluid
S. Mark Wainwright, Cláudia C. Mendes, Aashika Sekar, Benjamin Kroeger, Josephine E.E.U. Hellberg, Shih-Jung Fan, Abigail Pavey, Pauline Marie, Aaron Leiblich, Carina Gandy, Laura Corrigan, Rachel Patel, Stuart Wigby, John F. Morris, Deborah C.I. Goberdhan, Clive Wilson
Live-cell 3D single-molecule tracking reveals how NuRD modulates enhancer dynamics
S Basu, O Shukron, A Ponjavic, P Parruto, W Boucher, W Zhang, N Reynolds, D Lando, D Shah, LH Sober, A Jartseva, R Ragheb, J Cramard, R Floyd, G Brown, K Gor, J Balmer, TA Drury, AR Carr, L-M Needham, A Aubert, G Communie, L Morey, E Blanco, MA Barber, I Mohorianu, T Bartke, L Di Croce, I Berger, C Schaffitzel, SF Lee, TJ Stevens, D Klenerman, BD Hendrich, D Holcman, ED Laue
Adipose-tissue derived signals control bone remodelling
He Fu, Maria-Bernadette Madel, Dominique D. Pierroz, Mariano Schiffrin, Carine Winkler, Anne Wilson, Cécile Pochon, Barbara Toffoli, Jean-Yves Jouzeau, Federica Gilardi, Serge Ferrari, Nicolas Bonnet, Claudine Blin-Wakkach, Béatrice Desvergne, David Moulin
Fluorogenic probe for fast 3D whole-cell DNA-PAINT
Kenny Kwok Hin Chung, Zhao Zhang, Phylicia Kidd, Yongdeng Zhang, Nathan D Williams, Bennett Rollins, Yang Yang, Chenxiang Lin, David Baddeley, Joerg Bewersdorf
Cell division in the early C. elegans embryo. Plasma membrane, green; chromosomes, magenta.
Project Description
Cell division is fundamental to life and errors can result in abnormal chromosomal numbers, developmental defects, and cancers. Similarities in the structural and molecular organization of the division apparatus gives the textbook picture, that mechanisms underlying division, including formation and constriction of an actomyosin contractile ring, do not vary between cell and organism types. However, recent research and clinical findings suggest that there is a previously un-appreciated variation in the molecular requirement for cytokinesis to occur, as depletion or mutation of several ‘essential’ proteins only disrupt division in specific cell-types and lineages.
During C. elegans development each cell has a specific identity and gives rise to different cell lineages. Furthermore, the factors that control cell identity have been well mapped by decades of detailed developmental biology. Therefore, it is an excellent multicellular system for observing the interaction between cell identity and cytokinetic perturbation in genetically identical, but functionally variable cells.
This PhD project will use the C. elegans model system and state of the art confocal microscopy to image cytokinesis, tracking the outcome of this rapid and dynamic cellular process, and quantifying localisation of fluorescently labelled cytokinetic proteins, while perturbing different aspects of cell identity. During the PhD, the student will receive training in live-cell fluorescence microscopy and image analysis, as well as general molecular and cell biology techniques and C. elegans genetics and transgenesis (via MosSCI and CRISPR). Furthermore, they will have the opportunity to take part in DTP-wide training and networking events, an external research placement, and attend conferences.
NOTE: Application details are below, however it is strongly recommended to contact the primary supervisor prior to applying.
Applications should be made by emailing bbsrcdtp@liverpool.ac.uk with a CV (including contact details of at least two academic (or other relevant) referees), and a covering letter – including whatever additional information you feel is pertinent to your application; you may wish to indicate, for example, why you are particularly interested in the selected project and at the selected University. Applications not meeting these criteria will be rejected.
Please note that the closing date for applications is Monday 18th May at 12noon.
Funding Notes
This is a 4 year BBSRC studentship under the Newcastle-Liverpool-Durham DTP. The successful applicant will receive research costs, tuition fees and stipend (£15,009 for 2019-20). The PhD will start in October 2020. Applicants should have, or be expecting to receive, a 2.1 Hons degree (or equivalent) in a relevant subject. EU candidates must have been resident in the UK for 3 years in order to receive full support. Please note, there are 2 stages to the application process.
2) FLIRT: Fast Local InfraRed Thermogenetics for subcellular control of protein function. 2018 Nature Methods (15) 921–923
3) Using fast-acting temperature sensitive temperature sensitive mutants to study cell division in Caenorhabditis elegans. 2017 Methods in Cell Biology (137) 283-306
4) Cortical PAR polarity proteins promote robust cytokinesis during asymmetric cell division. 2016 Journal of Cell Biology (212) 39 – 49
5) High-Resolution Temporal Analysis Reveals a Functional Timeline for the Molecular Regulation of Cytokinesis. 2014 Developmental Cell (30) 209 – 223
6) aPKC cycles between functionally distinct PAR protein assemblies to drive cell polarity. 2017 Developmental Cell (4):400-415.e9
In the early fly embryo, information encoded in a handful of maternally deposited protein gradients is fed forward through increasingly intricate layers of interacting genes, culminating in the differentiation of the embryo into functional body segments with a high degree of spatial and temporal precision[1],[2]. Short segments of regulatory DNA known as enhancers lie at the heart of this developmental cascade. Enhancers contain binding sites for transcription factor proteins and are thought to act like computational units that “read” input concentrations of relevant transcription factors and “compute” a corresponding level of output gene expression[3]. But while we know quite a bit about what enhancers do, the how remains a mystery: we still lack a quantitative physical understanding of the chain of molecular events that connect transcription factor binding at enhancer DNA to the activation or inhibition of transcription.
Our recent work was aimed at filling this gap through a combination of live imaging experiments and theoretical modeling that allowed us to investigate transcriptional regulation not only across space, but also over time[4]. We focused on the even-skipped (eve) stripe 2 enhancer, which drives a sharp stripe of pattern formation in developing embryos of the fruit fly Drosophila melanogaster. Nearly three decades ago, a seminal study by Steve Small and Michael Levine established the basic principles of eve stripe 2 regulation: two activators, Hunchback and Bicoid, initially establish a broad domain of expression that is then refined into a sharp stripe by the action of the repressors Giant and Krüppel on the anterior and posterior sides of the stripe, respectively[5]. Dozens of papers have followed in their footsteps, making eve stripe 2 one of the most widely studied enhancers in developmental biology.
In 2014, the dynamics of eve stripe 2 expression in living embryos were examined for the first time using the MS2 system[6],[7]. These experiments revealed that the rate of transcription at individual eve loci was highly stochastic, with periodic “bursts” of rapid transcription that were separated by periods with little to no activity. This transcriptional bursting has been observed across a wide variety of model organisms, suggesting that it is the rule, rather than the exception[8],[9],[10],[11],[12],[13]. Because these single-locus fluctuations reflect the molecular processes that drive transcription, the authors speculated that it might be possible to learn more about the molecular mechanisms that underpin transcriptional regulation by taking a quantitative look at how the characteristics of transcriptional bursts varied across different regions of the eve stripe 2 pattern. Yet at that point there was no rigorous way to quantify how transcriptional bursting changed as a function of space and time. The challenge we faced was that rather than reporting on the instantaneous state of the promoter, the fluorescence readout from MS2 experiments at each time point corresponded to the aggregate promoter activity coming from all RNA polymerases actively transcribing the gene and elongating nascent RNA (see Figure 1C). To overcome this challenge, we worked with Hernan Garcia and Chris Wiggins to develop a new computational tool that could systematically dissect burst dynamics at individual loci across a pattern of gene expression.
cpHMM inference
Coined “compound state hidden Markov model” (cpHMM), our new computational technique is conceptually similar to an approach that was recently developed in the Chubb lab[14] and earlier HMM-based approaches[15],[16]. It allowed us to deconvolve the MS2 traces and obtain information about the instantaneous promoter activity at individual eve loci, which we described using a simple “telegraph” model with one active and one inactive state (Figure 1A & C, Video 1). This model has three parameters: the kinetic on- and off-rates that control the switching between transcriptionally active and inactive promoter states, and the rate of RNA polymerase initiation. Thus, in the model, there are three different “knobs” that transcription factors can tune to increase or decrease the average rate of mRNA production at a locus. With this technique in hand, we then set out to infer how these parameters were controlled across space and time from individual MS2 traces obtained from live imaging.
Our inference results showed that regulation of bursting took place primarily through the spatiotemporal modulation of on-rates (bursting frequency), with eve loci in the stripe center bursting with about twice the frequency of those at the stripe boundaries (Figure 1B). This finding suggests that the transcription factors responsible for modulating eve activity act by speeding up or slowing down one or more of the molecular steps leading to the initiation of transcriptional bursts, but do not affect the rate of RNA polymerase loading within a burst or the duration of the burst (inverse of the off-rate). Thus, although we did not have single-molecule resolution, we were able to use our computational method to infer the transcriptional state of individual promoters and to make headway towards a molecular understanding of the mechanisms driving transcriptional control.
Figure 1: cpHMM inference based on a kinetic model of promoter activity. (A) Random telegraph model of gene activity with three kinetic rate parameters. (B) Inferred spatial modulation of the burst frequency. (C) MS2 traces, representing aggregate signals from all elongating polymerases, deconvolved to yield the hidden promoter state dynamics.
Temporal, not spatial regulation explains the majority of stripe formation
Our cpHMM inference provided a prediction for the shape of the eve stripe 2 pattern, assuming that it was formed entirely through the spatial control of the burst frequency (green profile in Figure 2B). To our surprise, this “bursting only” prediction significantly underestimated the true dynamic range of the stripe pattern as revealed by our live imaging data (red profile in Figure 2B).
Going back to the raw data, we realized what was happening: the most striking aspect of eve stripe 2 expression is not the control of the average rate of mRNA production (via the control of transcriptional bursting), but, instead, the simple fact that the period of time over which gene loci engaged in bursty transcription was sharply controlled across the stripe. Eve loci in the stripe center were active for >40 minutes, while those on the stripe flanks were active for only 10-15 minutes. This indicated that the stripe pattern was being driven by the joint action of two distinct regulatory strategies, the control of the mean rate and the control of the transcriptional time window (Figure 2A).
We developed a simple quantitative model to connect observed control of the transcriptional time window activity over time with the predicted corresponding pattern of eve mRNA concentration. The model indicated that, indeed, this simple control of the duration of the period of transcriptional activity amongst nuclei on the stripe flanks accounted for the majority of stripe formation (blue profile, Figure 2A), while the control of the rate of mRNA production during the active window–the modality that had been the focus of most studies to date–only served to refine and sharpen this stripe. Thus, while transcriptional bursting certainly provides a window into the molecular nature of transcriptional control, the regulation of the period of time over which transcription occurs is the key strategy employed by the fly to realize the sharp eve stripe 2.
Figure 2: Two regulatory strategies driving stripe formation. (A) Our analysis revealed that eve stripe 2 is generated through the interplay between two different regulatory strategies: the control of the transcription rate while loci are active, and the control of the amount of time that loci are active. (B) A simple model indicated that the control of the transcriptional time window plays the dominant role in driving stripe formation.
We hypothesized that these two “control strategies” for stripe formation (control of average mRNA production rate and control of the transcriptional time window) might result from two different underlying molecular mechanisms. To test this hypothesis, we developed a simple model that used logistic regressions to relate the fraction of nuclei that had ceased transcribing at a given time and location to the concentrations of the four known eve stripe 2 regulators. We applied this framework to our live imaging data, along with a time series of transcription factor concentrations derived from fixed tissue experiments that was previously published by the Gregor Lab at Princeton[17]. The results indicated that the timing with which eve loci on the stripe boundaries stopped transcribing could be explained entirely by the progressive increase in the levels of Giant and Krüppel over time, and–surprisingly–was insensitive to the concentration dynamics of the activators Bicoid and Hunchback. These results suggest that the repressors act to turn off eve loci in nuclei on the stripe edges via a molecular pathway that is orthogonal to the one that controls transcriptional bursting and the mean rate of transcription.
Video 1: Decoding instantaneous promoter state dynamics inferred from MS2 imaging using the cpHMM method. Green and red nuclear coloring corresponds to ON and OFF promoter states in transcriptionally engaged loci, respectively.
Future directions
While our study focused on the pattern driven by the eve stripe 2 enhancer in developing Drosophila embryos, the quantitative techniques we developed to dissect transcriptional bursting and to connect output transcription dynamics to input transcription factor concentrations are quite general. It is our hope that these tools can be used by researchers to gain insight into the transcriptional regulation of other genes both in Drosophila and in other organisms. To facilitate this, we are making the code for cpHMM publicly available at https://github.com/GarciaLab/cpHMM. With respect to new research directions, we have recently collaborated with the Eisen Lab at Berkeley to study the transcriptional dynamics of the full eve locus. There, we found that despite being created by the largely independent activity of five discrete enhancers, the seven eve stripes are sculpted by the same basic regulatory strategies[18].
In addition to inferring the kinetic parameters of promoter activity, our methodology also allows us to decode individual activity traces and find the most likely promoter state sequences (Figure 1C, Video 1). In future work, we hope to utilize this feature of our methodology to dissect the temporal interplay between the fluctuations in the local concentration of transcription factors at a transcriptional locus and the initiation of transcriptional bursts. Specifically, we plan to correlate the inferred promoter state sequence information with real-time measurements of transcription factor concentrations in individual nuclei, which is now possible due to the LlamaTag technology that was recently developed by our colleague Jacques Bothma[19]. Combining these computational and live imaging techniques will allow us to examine the concentration of transcription factors at the start and end of transcriptional bursts, shedding light on how individual transcription factors act within the transcriptional cycle.
Written by Nick Lammers, Vahe Galstyan and Hernan Garcia
(28 votes)
Loading...
In the latest episode of Genetics Unzipped,
(No Ratings Yet)
(15 votes)
:
