Some recent papers by R P Futrelle - For NSF Proposal of 11/2002

Note: The copyright to most of these papers is held by the original publisher.


Futrelle, R. P. & Rumshisky, A. (2001). "Discourse Structure of Text-Graphics Documents" 1st International Symposium on Smart Graphics Hawthorne, NY. ACM Press. [Link to paper]

Futrelle, R. P. (1999). "Summarization of Diagrams in Documents" In I. Mani & M. Maybury (Eds.), Advances in Automated Text Summarization. Cambridge, MA: MIT Press. [Link to paper]

Futrelle, R. P. (1999). "Ambiguity in Visual Language Theory and its Role in Diagram Parsing" IEEE Symposium on Visual Languages, VL99. IEEE Computer Soc., Tokyo, pp. 172-175. [Link to paper]

Futrelle, R. P., and N. Nikolakis. (1995) "Efficient Analysis of Complex Diagrams using Constraint-Based Parsing" 782-790, ICDAR-95 (Intl. Conf. on Document Analysis & Recognition), Montreal, Canada. pp. 782-790. [Link to paper]

Gauch, S., and R. P. Futrelle. 1993. "Experiments in Automatic Word Class and Word Sense Identification for Information Retrieval" Third Annual Symposium on Document Analysis and Information Retrieval. pp. 425-434. [Link to paper]


Futrelle, R. P., and N. Fridman. (1995) "Principles and Tools for Authoring Knowledge-Rich Documents" DEXA 95 (Database and Expert System Applications) Workshop on Digital Libraries, London, UK pp. 357-362. [Link to paper]

Futrelle, R. P. (1998) The Diagram Understanding System Demonstration Site

Futrelle, R. P., J. Traut and W. G. McKee . 1982. "Cell behavior in Dictyostelium discoideum: Preaggregation response to localized cyclic AMP pulses" J. Cell Biology, (92), 807-821. Link to paper

Futrelle, R. P., R. J. Williams, W. Tong, and M. J. Ondrechen. (2002). "Automating the prediction of enzyme active sites from structure alone in THEMATICS" Poster Presented at the 224th ACS National Meeting (American Chemical Society), Boston, MA, August 18-22, 2002. Abstract immediately below.

THEMATICS is a method that uses theoretical microscopic titration curves to identify active-site residues in proteins of known structure. We have shown that predicted titration curves that do not have the typical Henderson-Hasselbalch shape are significant for proteins; they are markers of active site location and chemical reactivity. We are developing machine-learning methods to perform active site identification automatically. The training data consist of titration curves computed for all of the ionizable residues of each protein where the active site has already been well characterized by biochemists. In the initial studies, titration curve parameters along with relative locations are used as inputs to neural network algorithms. The parameters include effective pKa, maximal slope, and the pH range over which the curve separates from its asymptotes. Estimates are presented of performance on as yet unseen enzymes (cross-validation). Lessons learned from these initial studies are described.