No abstract

Three-dimensional (3-D) structural models of RNA are essential for understanding of the cellular roles played by RNA. Such models have been obtained by a technique based on a constraint satisfaction algorithm that allows for the facile incorporation of secondary and other structural information. The program generates 3-D structures of RNA with atomic-level resolution that can be refined by numerical techniques such as energy minimization. The precision of this technique was evaluated by comparing predicted transfer RNA loop and RNA pseudoknot structures with known or consensus structures. The root-mean-square deviation (2.0 to 3.0 angstroms before minimization) between predicted and control structures reveal this system to be an effective method in modeling RNA.

The problem of choosing an alignment of two or more nucleotide sequences is particularly difficult for nucleic acids, such as 5S ribosomal RNA, which do not code for protein and for which secondary structure is unknown. Given a set of 'costs' for the various types of replacement mutations and for base insertion or deletion, we present a dynamic programming algorithm which finds the optimal (least costly) alignment for a set of N sequences simultaneously, where each sequence is associated with one of the N tips of a given evolutionary tree. Concurrently, protosequences are constructed corresponding to the ancestral nodes of the tree. A version of this algorithm, modified to be computationally feasible, is implemented to align the sequences of 5S RNA from nine organisms. Complete sets of alignments and protosequence reconstructions are done for a large number of different configurations of mutation costs. Examination of the family of curbes of total replacements inferred versus the ratio of transitions/transversions inferred, each curve corresponding to a given number of insertions-deletions inferred, provides a method for estimating relative costs and relative frequencies for these different types of mutations.

This work is in the context of TRANSTYPE, a system that observes its user as he or she types a translation and repeatedly suggests completions for the text already entered. The user may either accept, modify, or ignore these suggestions. We describe the design, implementation, and performance of a prototype which suggests completions of units of texts that are longer than one word.

We present and evaluate SumUM, a text summarization system that takes a raw technical text as input and produces an indicative informative summary. The indicative part of the summary identifies the topics of the document, and the informative part elaborates on some of these topics according to the reader's interest. SumUM motivates the topics, describes entities, and defines concepts. It is a first step for exploring the issue of dynamic summarization. This is accomplished through a process of shallow syntactic and semantic analysis, concept identification, and text regeneration. Our method was developed through the study of a corpus of abstracts written by professional abstractors. Relying on human judgment, we have evaluated indicativeness, informativeness, and text acceptability of the automatic summaries. The results thus far indicate good performance when compared with other summarization technologies.

Text prediction is a form of interactive machine translation that is well suited to skilled translators. In principle it can assist in the production of a target text with minimal disruption to a translator's normal routine. However, recent evaluations of a prototype prediction system showed that it significantly decreased the productivity of most translators who used it. In this paper, we analyze the reasons for this and propose a solution which consists in seeking predictions that maximize the expected benefit to the translator, rather than just trying to anticipate some amount of upcoming text. Using a model of a "typical translator" constructed from data collected in the evaluations of the prediction prototype, we show that this approach has the potential to turn text prediction into a help rather than a hindrance to a translator.

The convergence of ancestral sequences independently constructed from different branches of a phylogenetic tree can be used as a test of homology of data sequences. This criterion has shown that all phenylalanine tRNAs are related to a common ancestor, whereas eukaryotic and prokaryotic tyrosine tRNAs may have independent origins. All glycine tRNAs share a common ancestor. The glycine tRNA family splits according to the purine or pyrimidine nature of the first anticodon base prior to the divergence of eukaryotes and prokaryotes. The structural similarity between some prokaryotic glycine and valine tRNAs is the result of their derivation from a common ancestor that existed previous to the divergence of the different glycine tRNAs. These results support models of genetic code evolution involving the incremental elaboration of earlier, simpler codes.

No abstract