Getting Started with RA: What are Claims? How do you work with them?

By Scott Needham

Posted in How to use Research Analysis and FAQ

The Research Analysis platform is centred around the concept of the scientific claim. The goal of a claim is to take a hypothesis that has been argued in the literature and state it in a formalised language that allows for easier comparison, searching and analysis.

While there are several types of scientific claim, Research Analysis is currently focused on Cause-Effect type scientific claims eg. Statins Decrease Coronary Artery Disease in Humans. We have broken up the components of Cause-Effect claims and formalised their language and structure to make them easier to compare, search and analyse. The easiest way to see how they work is to look through some examples in the View Claims table, but here we provide an overview of the model.

Figure 1. Research Analysis Knowledge Management Model

• Claim Elements:
  1. Treatment/Cause: This is the drug, environmental, genetic, etc cause that is made in the model system. eg. Statin treatment
  2. Effect: This is the type of effect that results from the Treatment/Cause. Currently we only offer three options: increases, decreases or not significant.
  3. Molecule/Disease: This describes the molecule or disease that is effected by the cause. eg. Cholesterol
  4. Organ Model: This describes the organ of the animal or the in vitro system where the cause-effect relationship was observed. eg. Blood
  5. Genetic Model: This describes the genetic model in which the cause-effect relationship was observed. eg. ApoE -/-, Wild Type.
  6. Species: This describes the species in which the cause-effect relationship was observed. eg. Human.
• Standard Terms: Wherever possible, claims should use standardised language for the elements of the claims. Research Analysis currently requires that all terms should be sourced from one of the following, in order:
  • Medical Subject Headings (MeSH) standard – www.nlm.nih.gov/mesh. Elements of claims should either exist in MeSH as Subject Headings or any of the Entry Terms for a heading or the Tree Number/Unique ID. For example, Humans is a Subject Heading in MeSH, but you can also use the Entry Terms which include Human, Homo sapiens, Man.
  • NCBI Protein Database – www.ncbi.nlm.nih.gov/protein.The name including synonyms or the unique code eg. Accession for the relevant database can be used in Research Analysis.
  • NCBI PubChem Database – www.ncbi.nlm.nih.gov/protein. The chemical name including synonyms or the PubChem CID code can be used in Research Analysis.

We do allow for new terms to be added where there is no satisfactory term in the above databases, however the use of standardised terms allows for much more powerful search and analysis. For example, a scientist in one field may use a different term to those in another field, but using standard terms we can link these two claims together and provide cross-field visibility. In future we also hope to provide searching and analysis across different levels of the term tree eg. Hominids or Mammals would consolidate claims for species further out on the tree.

• Supporting Quotes: Research Analysis requires that there be a quoted statement taken from the literature that supports each scientific claim in the database. While the whole research paper is required to fully support a claim, the quoted statements provide some context to the claim in the words of the scientist. It also makes it easier for other scientists to identify the section of the paper that was used to create the formalised claim.
• Referencing: Along with the Supporting Quote, Research Analysis requires that the PMID is provided for the paper that the Supporting Quote was taken from – www.ncbi.nlm.nih.gov/pubmed. This allows users to quickly dive into the paper for more information or context by clicking on the PMID link in Research Analysis.

How to capture claims using Research Analysis?

• Claims to Capture: A research paper may introduce one new claim, but generally presents a collection of claims that support a central claim of the paper. The user can choose to add just the highest level claim of the paper or drill down and capture the supporting claims too. It is preferred that claims are entered only for papers that provide new evidence or verification of the claim and not for papers that simply refer to the findings of other papers. Ideally you would drill down into the paper referenced and enter the claim with a quote from this paper. For this reason we prefer that claims are entered from original research papers or papers that verify previous work, rather than from review papers.
• Adding Claims: You can add claims to the database via two means:

1. Add Claim on View Claims page: To enter a claim on this page you simply need to enter the values into each of the boxes at the top of the table and then press the “Add Claim” button. The nice thing about this method is that it simultaneously searches the database to see if the claim already exists and you’ll see related claims as you enter each of the terms.
2. Upload a Spreadsheet of Claims: To load a batch of claims you can go to the Upload Claims menu option and select and upload a spreadsheet containing claims. The spreadsheet will need to be in the right format and you can learn more in this article.

• Edit/Delete Detailed Claim View: To edit, delete or see more details relating to a claim you can click the View link in the View Claims table and go to the claim page. Where the claim is supported by several Quoted Statements and papers you can see a listing of these on the claim page. You can also get a unique link to the claim page that makes it easy to share the claim with other users of Research Analysis.

We hope that this helps you get started with Research Analysis, but please feel free to contact us if you have further questions.