Papers by Aleksandar Radovanovic
PLoS ONE, 2013
Background: Sickle cell disease (SCD) is a fatal monogenic disorder with no effective cure and th... more Background: Sickle cell disease (SCD) is a fatal monogenic disorder with no effective cure and thus high rates of morbidity and sequelae. Efforts toward discovery of disease modifying drugs and curative strategies can be augmented by leveraging the plethora of information contained in available biomedical literature. To facilitate research in this direction we have developed a resource, Dragon Exploration System for Sickle Cell Disease (DESSCD) () that aims to promote the easy exploration of SCD-related data. The Dragon Exploration System (DES), developed based on text mining and complemented by data mining, processed 419,612 MEDLINE abstracts retrieved from a PubMed query using SCD-related keywords. The processed SCDrelated data has been made available via the DESSCD web query interface that enables: a/information retrieval using specified concepts, keywords and phrases, and b/the generation of inferred association networks and hypotheses. The usefulness of the system is demonstrated by: a/reproducing a known scientific fact, the ''Sickle_Cell_Anemia-Hydroxyurea'' association, and b/generating novel and plausible ''Sickle_Cell_Anemia-Hydroxyfasudil'' hypothesis. A PCT patent (PCT/ US12/55042) has been filed for the latter drug repurposing for SCD treatment. We developed the DESSCD resource dedicated to exploration of text-mined and data-mined information about SCD. No similar SCD-related resource exists. Thus, we anticipate that DESSCD will serve as a valuable tool for physicians and researchers interested in SCD.
Infection, Genetics and Evolution, 2011
Even though Hepatitis C Virus (HCV) cDNA was characterized about 20 years ago, there is insuffici... more Even though Hepatitis C Virus (HCV) cDNA was characterized about 20 years ago, there is insufficient understanding of the molecular etiology underlying HCV infections. Current global rates of infection and its increasingly chronic character are causes of concern for health policy experts. Vast amount of data accumulated from biochemical, genomic, proteomic, and other biological analyses allows for novel insights into the HCV viral structure, life cycle and functions of its proteins. Biomedical text-mining is a useful approach for analyzing the increasing corpus of published scientific literature on HCV. We report here the first comprehensive HCV customized biomedical text-mining based online web resource, dragon exploratory system on Hepatitis C Virus (DESHCV), a biomedical text-mining and relationship exploring knowledgebase was developed by exploring literature on HCV. The pre-compiled dictionaries existing in the dragon exploratory system (DES) were enriched with biomedical concepts pertaining to HCV proteins, their name variants and symbols to make it suitable for targeted information exploration and knowledge extraction as focused on HCV. A list of 32,895 abstracts retrieved via PubMed database using specific keywords searches related to HCV were processed based on concept recognition of terms from several dictionaries. The web query interface enables retrieval of information using specified concepts, keywords and phrases, generating text-derived association networks and hypotheses, which could be tested to identify potentially novel relationship between different concepts. Such an approach could also augment efforts in the search for diagnostic or even therapeutic targets. DESHCV thus represents online literature-based discovery resource freely accessible for academic and non-profit users via and its mirror site .
Rice, 2008
Coordinated transcriptional modulation of large gene sets depends on the combinatorial use of cis... more Coordinated transcriptional modulation of large gene sets depends on the combinatorial use of cis-regulatory motifs in promoters. We postulate that promoter content similarities are diagnostic for co-expressing genes that function coherently during specific cellular responses. To find the coexpressing genes we propose an ab initio method that identifies motif families in promoters of target gene groups, map these families to the promoters of all genes in the genome, and determine the best matches of each of the target group gene promoters with all other promoters. When the method was tested in rice starting from a group of co-expressing Late Embryogenesis Abundant (LEA) genes, we obtained a promoter similarity-based network that contained candidate genes that could plausibly complement the function of LEA genes. Importantly, 73.36% of 244 genes predicted by our method were experimentally confirmed to co-express with the LEA genes in maturing rice embryos, making this methodology a promising tool for biological systems analyses.
Uploads
Papers by Aleksandar Radovanovic