Hackathons (or “code fests”) are a fun away to bring together diverse audiences to synergize ideas for problem solving. Recent biomedical Hackathons have led to a number of data science tools that became widely utilized and subsequently gained independent funding.
We are hosting a biomedical Hackathon that explicitly encourages the participation of ISHR investigator community, not just software developers or programmers. In the past, Hackathons participated by software engineers have been productive in prototyping algorithm solutions for well-defined computational tasks. In this particular Hackathon, we would like to encourage the participation of biomedical investigators and capitalize on their expertise regarding use cases and provide valuable scientific contexts for spontaneous projects.
The overarching theme of our Heart-bits Hackathon is disease signature discovery from phenotyping data. We anticipate the participants will gain experience in generating new ideas and new hypotheses (ideathon) by a data science approach.
Date: At the ISHR World Congress (April 20, 2016)
Venue: Buenos Aires, Argentina
Room: 204, 2nd floor of the UCA convention center
- ISHR investigators (faculty and trainees)
- Hackers, Techies, and Gurus
To reach a broad audience, we have designed the event to comprise both a “General Audience Track” and a “Hacking Track”. The General Track will include a one-hour "Big Data Introduction" session and a "Project Presentation" session. The Hacking Track will include a "Machine Learning Practicum" session and a "On-the-Cloud Collaborative Analysis" session.
1. Big Data Introduction - (general track) A one-hour introduction to data science and its roles in biomedical research. Topics covered will include current big data themes, real-world big data use cases, and potential applications of big data to cardiovascular research, followed by an overview of our hacking sessions.
2. On-the-Cloud Collaborative Data Analysis - (hacking track) Participants will gain experiences on collaborative analysis using on-the-cloud software tools to extract new biomedical insights. As datasets grow larger, faster hardware and sophisticated computer architectures become essential to biomedical research. By performing data analysis on remote servers (called "cloud servers"), any research group will be able to tackle big data problems without making costly investments in computer hardware. In this hacking session, participants will be provided with a public omics dataset on an isoproterenol-stimulated mouse model of cardiac hypertrophy. The dataset will include experimental values on the turnover rates, expression, UniProt accessions and names of proteins. These data will be analyzed using on-the-cloud software tools including Sage Synapse that provides a online collaborative platform, Cytoscape and Reactome that enable collaborative network analysis, as well as COPaKB and Cardiac GeneWiki that offer cardiac specific knowledge on proteins. Participants are also free to explore other creative ideas for software tool development and collaborative analysis of disease signatures, pathways, or networks.
3. Machine Learning Practicum - (hacking track) Participants will gain hands on experience with machine learning package algorithms, following a tutorial for beginners. Machine learning uses small datasets to build computational model that predicts the outcome of a user-defined question. Here is an example use case: the participants will be provided with an echocardiogram dataset on patients who suffered heart attacks. The data will include values on the patients’ age at first heart attack, presence of pericardial-effusion, fractional-shortening, E-point septal separation, left ventricular end-diastolic dimension, left ventricular wall motion score and left ventricular wall motion index. These data will be used to train the computational prediction model. Using a logistic regression algorithm in the Octave programming language, the participants will predict the survival time course of a patient who experienced a heart attack. Through this hands-on session, new hackers will be exposed to general concepts of machine learning (e.g., cost function and feature mapping) that are broadly applicable to cardiovascular research.
4. Project Summary and Presentation - (general track) Hacking teams will take turns to present their ideas and findings from each of the hacking sessions. Audiences will be provided an online Google form to vote for their favorite teams. One travel award will be given to one presenting ECI (early career investigator) from each of the 3 teams with the greatest number of votes.
A total of 3 Travel Awards are available for ECIs (early career investigators) selected during the Project Summary and Presentation section:
- Each award will for up to $500 and will be given to support travel expenses.
- Awardee must participate in all sessions of the Hackathon.
Please sign up at https://goo.gl/zEj4hs
Questions or Comments
Please send any comments or questions to Dr. Maggie Lam at firstname.lastname@example.org