Gianluca Pollastri


I am an Associate Professor in the School of Computer Science at University College Dublin. Since the beginning of 2007 I am based at the Institute for Discovery. I have worked as a PI in Dublin since September 2003. Before then I was a working in Pierre Baldi's group at University of California, Irvine, where it was a little bit drier than in Dublin, but Guinness tasted remarkably foul.
I am the head of a lab, which has numbered as many as two postdoctoral fellows and ten postgraduate students, and has been funded through a number of grants from Science Foundation Ireland, the Health Research Board of Ireland, the Irish Research Council for Science, Engineering and Technology (IRCSET) and UC Dublin.
I am an Erdös number 3.

Research interests

There are two main related focuses to my research: one is the development of Machine Learning models for complex structured data; the other is the fields of Bioinformatics and Chemoinformatics. The two things are inextricably tied, since virtually all the applicative domains I tested my models on are in molecular biology.
To get an idea of our ongoing projects, you may check our research blurb, the list of our free tools and servers, or our funding and publication pages.

Past Projects

The SCRATCH servers

Servers for the prediction of various structural features of proteins based on bidirectional recurrent neural networks (BRNNs), and 2D-Recursive Neural Networks. I developed BRNNs at the University of Florence and 2D-RNN at UC Irvine, and designed and built most of the SCRATCH servers while I was at University of California, Irvine. Although I have not worked on the servers since 2003, and strangely enough have not contributed to the paper describing the bundle of them, you can find a list of past SCRATCH literature here (papers 17 to 3).

Contact information

Office: E0.95, UCD Institute for Discovery,
Science East, UCD
Belfield, Dublin 4
Post: UCD Institute for Discovery
University College Dublin
Belfield, Dublin 4
email: gianluca.pollastri at
phone: ++353 1 716 2620