Search

There is increasing interest in and need for innovative solutions to medical search. In this paper we present the EU-funded Khresmoi medical search and access system, currently in year 3 of 4 of development across 12 partners. The Khresmoi system uses a component-based architecture housed in the cloud to allow for the development of several innovative applications to support target users' medical information needs. The Khresmoi search systems based on this architecture have been designed to support the multilingual and multimodal information needs of three target groups: the general public, general practitioners and consultant radiologists. In this paper we focus on the presentation of the systems to support the latter two groups using semantic, multilingual text and image-based (including 2D and 3D radiology images) search.

The Khresmoi project is developing a multilingual multimodal search and access system for medical and health information and documents. This scientific demonstration presents the current state of the Khresmoi integrated system, which includes components for text and image annotation, semantic search, search by image similarity and machine translation. The flexibility in adapting the system to varying requirements for different types of medical information search is demonstrated through two instantiations of the system, one aimed at medical professionals in general and the second aimed at radiologists. The key innovations of the Khresmoi system are the integration of multiple software components in a flexible scalable medical search system, the use of annotation cycles including manual correction to improve semantic search, and the possibility to do large scale visual similarity search on 2D and 3D (CT, MR) medical images.

Khresmoi is a European Integrated Project developing a multilingual multimodal search and access system for medical and health information and documents. It addresses the challenges of searching through huge amounts of medical data, including general medical information available on the internet, as well as radiology data in hospital archives. It is developing novel semantic search and visual search techniques for the medical domain. At the MIE Village of the Future, Khresmoi proposes to have two interactive demonstrations of the system under development, as well as an overview oral presentation and potentially some poster presentations.

Le transport de neige par le vent est un phénomène complexe où interviennent l'écoulement turbulent d'un fluide (le vent) et des particules solides (la neige). Ces deux composantes inter-agissent, le vent érodant et déposant la neige et, réciproquement, la forme du dépôt influençant le vent. Ce phénomène est riche au niveau théorique autant que pratique : la formation de corniches ou de congères peuvent menacer des habitations ou fermer des routes alpines. Il est alors nécessaire de prévoir et, par exemple, de tenter de lutter contre par la construction d'ouvrages. Des modèles, basés sur des solveurs de mécanique des fluides classiques ont été imaginés. Cependant, une autre approche existe, particulièrement bien adaptée aux écoulements complexes de fluide : il s'agit des modèles de gaz sur réseaux. Un tel modèle, original, a été développé et implémenté efficacement sur des ordinateurs parallèles, ouvrant la voie à des nouvelles perspectives.

Khresmoi is a European Integrated Project developing a multilingual multimodal search and access system for medical and health information and documents. It addresses the challenges of searching through huge amounts of medical data, including general medical information available on the internet, as well as radiology data in hospital archives. It is developing novel semantic search and visual search techniques for the medical domain. At the MIE Village of the Future, Khresmoi proposes to have two interactive demonstrations of the system under development, as well as an overview oral presentation and potentially some poster presentations.

LLC-MSMS combined with library search approaches is getting popular as an alternative to existing methods in clinical toxicology, for the screening of toxics. Such screenings are traditionally operated in emergency units, where they act as an early diagnostic test of the patient condition. As a result, routine laboratories need robust and easy to use software solutions that can handle the complexity of spectra generated by LCMSMS as well as deliver high quality results in the shortest time possible. In this thesis, we present a small molecule identification platform called SmileMS. This platform was developed to be used in applications based on LC-MSMS identification but also to act as an exploration platform for further investigation in the field of small molecule identification. Throughout the conception of SmileMS, these investigations concerned four areas: the creation scoring models, the handling of spectral libraries, the theoretical fragmentation of molecules, and the setting of engineering strategies enabling software exploration of data.

Identification of small molecules is of major importance for many applications. Liquid Chromatography Tandem Mass spectrometry (LC- MSMS) is gaining increasing interest in the field of small molecule identification. LC-MSMS has a broad range of detection, is sensitive and does not need special sample pre-processing. As a major chal- lenge, spectra of the same compound can show great variability across acquisitions. High spectra variability limits the use of LC-MSMS for library search identifications. Dedicated identification tools such as MS Search from NIST show insufficient performances when it comes to cross-platform identification. In this thesis, we present the new library search scoring model X- Rank. X-Rank matches conserved properties of spectra and proposes a robust probability scoring model. Scoring parameters can be opti- mized from a training set. A re-training of X-Rank for a specific data set, was shown to essentially improve the results. The efficiency of X-Rank was compared to existing solutions, using two test-sets from different machine types. Overall X-Rank showed better results in terms of sensitivity and specificity. Especially in the case of cross-platform identification, X-Rank could better discriminate correct from wrong matches. Furthermore, X-Rank could correctly identify and top rank eight chemical compounds in a test mix. Even though these results confirm an important improvement for cross-platform identification, filters before and after the X-Rank scor- ing are still useful. In this perspective, a new approach to confidently use the retention time information is presented. Furthermore, a spec- tra filtering approach is applied, which improves the identification in terms of quality and speed. Finally, using a specific training configuration, X-Rank was adapted for proteomics data. Combined with the peptide identification tool Phenyx, X-Rank helped matching additional peptides. X-Rank was implemented into the small molecule identification platform SmileMS. SmileMS is designed for a routine use in laborato- ries. It is a multi-user platform, which provides a simple identification workflow and intuitive result visualization. Thanks to the generic software architecture and the mutual inte- gration with the open-source project Java Proteomic Library (JPL), the addition of new methods to SmileMS is facilitated. Such methods in- clude quantification, the combined use of several algorithms, GC-MS and exact mass identification.