First Place: Automated visualization and quantification of fiber orientation

Kyle Quinn, Postdoctoral Scholar, Biomedical Engineering, Tufts School of Engineering

Click on the image to view original.

Detecting patterns of fiber orientation are necessary in a variety of engineering and biomedical applications. Typically scanning electron microscopy (SEM), second harmonic generation (SHG) imaging, or tissue histology techniques are employed to visualize fiber organization at different scales. However, subtle patterns in fiber alignment within an image are often difficult to detect and quantify. We have developed a unique algorithm to rapidly quantify fiber orientation at each pixel in a variety of standard biomedical images. MATLAB-based software was written to enable users to load medical images and view color-coded images of the local fiber orientation. Additionally, users can define any region of interest within the image and the program will produce a histogram of the local orientation distribution and compute summary directional statistics. This software can be used to provide quantitative outcomes for any application in which fiber orientation or organization is relevant.

Second Place: A Software Tool for Visualizing, Curating, and Mining Regenerative Experiments

Daniel Lobo, Postdoctoral Associate, Biology, Tufts School of Arts and Sciences
Taylor J. Malone, Undergraduate Student, Biology, Tufts School of Arts and Sciences
Michael Levin, Vannevar Bush Professor, Biology, Tufts School of Arts and Sciences

Click on image to view original.

To download the software and run the dataset go to the following webiste:

Many living organisms have remarkable regenerative abilities. Planarian flatworms are an important model system for molecular-genetic and biophysical studies of anatomical repair because they have an outstanding capacity to regenerate any missing body part. Amputated middle segments of the worm always regenerate a new head (including brain, eyes, muscles, etc.) and a new tail at the correct ends. Indeed, a piece as small as 1/270th of the adult worm regenerates into a whole animal\! Understanding the mechanisms that orchestrate these regeneration capabilities is a fundamental interest in biology and will directly impact the areas of cancer and regenerative medicine, having also implications for the design of robust, fault-tolerant robotic systems. Using state-of-the-art tools of molecular cell biology, many scientists are currently attempting to discover the networks and pathways that govern regeneration in planaria. However, no comprehensive model of the regenerative process has been found yet, and it is now clear that computational tools are needed to mine this ever-increasing huge dataset. As part of our efforts to establish a new Bioinformatics of Shape, we have created Planform (Planarian formalization) – a software tool for visualizing, curating, and mining published data from planaria regenerative experiments in a formal and unambiguous way. The software is based on a novel mathematical graph language to describe biological morphologies and their experimental manipulations. Interacting with a user-friendly graphical interface, any researcher can easily visualize, curate, and search planarian regenerative experiments. Using Planform, we have curated a database containing more than 1,000 experiments from the main publications in the planarian literature. Importantly, the tool and database are based on a mathematical formalism understandable by computers, paving the way for novel artificial intelligence tools that will extract knowledge from this dataset. We are currently developing such tools, which will allow us to build the first algorithmic, constructive model of planarian regeneration.

The 5 functions of our software that we would like the committee to focus on are:

  1. Originality: this is the first system to formalize systematically morphological experiments with a mathematical abstraction understandable by computers.
  2. Effective Communication: the details of complex experiments are presented comprehensively in a single screen that integrates clearly the specific experimental procedure and its outcome.
  3. Visual Impact: the software automatically draws effective cartoon diagrams of the encoded morphologies and manipulations, allowing the user to comprehend rapidly the details and results of the experiments.
  4. Usability: although morphological data needs a complex abstraction, the software presents a friendly click-and-drag graphical interface that permits to encode new morphologies and manipulations effectively by any user without training.
  5. Versatility: the software allows any scientist to create personal databases of new experiments, as well as access our public database including more than 1,000 published planarian experiments. A search module permits to query the database for experiments, worm manipulations, regenerated morphologies, etc. according to any criterion, such as name, publication title or year, drug or RNAi names, or number of specific organs or regions in the morphology.