DeCypher
          Products    News    Support    Sales    Contact 
  bioinformatics and text mining solutions

Site Search:  

Section Topics:

Introduction to AKS
1. Powerful searching
2. Find specific sentences
3. Highlight relevant text
4. Analyze large document sets
5. Informative ranking
6. Author ranking
7. Focusing searches
8. Manage your project
9. Developer Kit
10. Relate biomedical concepts
11. Knowledge navigation
12. External databases links
Inquire about AKS

News:

Tera-Probe for sensitive, high-throughput oligo searches

Announcing the CodeQuest™ Biocomputing Workstation

GeneMatcher & BlastMachine Replacement Program

DeCypher in the Literature

Newsletter:

Subscribe to the TimeLogic newsletter for product information, special offers and technology whitepapers!

Signup

Send page to a colleague

AKS: Large Document Set Analysis

When documents are retrieved by MEDLINE, the user is left with an enormous list of references that is impossible to read manually. AKS provides powerful features to analyze the content of large document sets. It first extracts all biological entities that appear in these documents and assigns a score to each of them that expresses the relevance of this entity in terms of the specific search.

Example 7 Fig 1Fig 1. Lists generated of the most relevant chemical substances (above) and the most relevant genes after analyzing the 29,912 documents containing "Alzheimer's disease"

AKS does not only retrieve documents matching a query, but also extracts all the biological entities (chemicals, diseases and genes) assigning to each one a value proportional to its relevance (biological meaningfulness) in the retrieved document set.

A list of the most significant chemical substances and genes in the entire document set containing Alzheimer's disease (more than 29,000) is displayed in Fig 1. By reviewing the chemical list, we find the most studied drugs to treat the disease standing out among the rest (12 occurrences out of 20). Most of them are acetylcholinesterase (ACHE) inhibitors, an enzyme closely related to Alzheimer's disease. Congruently, ACHE also appears in the list of the genes most significantly related to Alzheimer's disease (and other well-known genes related to the disease).

We can easily perform a search to retrieve documents containing both Alzheimer's disease and ACHE and carry out an analysis of this document subset. In the analysis results we find that two ACHE-inhibitors shown above appear among the most relevant sentences (Fig 2). We can also check at the cooccurrences of biomedical concepts under this document set, focusing on the chemicals list. We can see at the top of the ranking not only donepezil, rivastigmine and tacrine (that we observed in the sentence list) but also other relevant drugs for Alzheimer's disease such as galantamine, huperzine, physostigmine and eptastigmine. (Fig. 3)

Example 7 Fig 2
Fig 2. List of the most relevant sentences for the documents containing "Alzheimer's disease" and the enzyme "Acetylicholinesterase". Most of them contain names of acetylcholinesterase inhibitors (donepezil and tacrine).
Example 7 Fig 3
Fig. 3 List of cooccurring chemicals in the list of documents where "Alzheimer's disease" and "acetylcholinesterase" cooccurr. Most of them are inhibitors of the acetylcholinesterase and this is used as a strategy in the treatment of the Alzheimer's Disease.

 

View the TimeLogic 2006 Products, including AKS (pdf)

Request more information or schedule a demonstration of AKS today!

 

 

TimeLogic is a brand of Active Motif, Inc. All text © 2006