Survival analysis

KNIME Analytics Platform
1

Can survival analysis / Cox proportional hazards regression be performed using Knime? 

2

Hi,

this is possible via Python scripting nodes : https://github.com/CamDavidsonPilon/lifelines

Or R nodes https://cran.r-project.org/web/packages/survival/index.html

 

Best regards

Fabien

3

Is possible to perform logrank test in survival analysis (Kaplan Meier node) in Knime without R or Python?

4

Hello,
I cannot find the survival analysis node in my Knime program . ¿Any could help me? Thanks

5

The user referred to python and R so maybe you like to go with these nodes

6

Hello KNIME community,

We’re conducting research where we would like to search scientific studies for any study containing Kaplan-Meier survival curves for a specific disease.
The problem is that this data is not included in the titles or abstracts of studies, but is in the full text. Is there a method where we could search PubMed or Scopus for full text containing Kaplan-Meier curves without having to screen thousands of studies that have the word survival in the title and/or abstract?
(i.e. a tool that could help us search for the word Kaplan-Meier or overall survival inside the full texts?)

Thanks in advance,
Baher Elezbawy

7

Maybe there is an API which allows searching for key terms even if they are not in the title or abstract?
Edit: sth like this?

8

hi @baherelezbawy
I don’t think it is possible to search the full texts via API or web interface, at least on PubMed, which is, as long as I know, a bibliographic database, not a fulltext one. Perhaps it would be helpful to do some very targeted research on PubMed, download the results as a csv file, then use KNIME for example to select the main authors and journals and focus on them. I’ve tried it myself, searching “Kaplan-Meier” in all fields, then refining the search using some filters, finally downloading the csv file (7313 rows). This is an example workflow where I rank authors and journals according to the number of articles.

immagine
immagine1101×504 75 KB

KNIME_project.knwf (146.5 KB)
Using the DOIs you can reach the fulltext. For example, the paper with DOI 10.2217/fon-2021-0101 can be found here: https://dx.doi.org/10.2217/fon-2021-0101

1 Like
10

Thank you so much Duristef

11

Thank you so much Daniel

12

On second thoughts, this is not a good idea, mainly because of the author’s name abbreviation (e.g. “Zhang Y”).
PubMed is a bibliographic database, but PubMed Central is fulltext and allows fulltext searching

immagine
immagine919×150 6.43 KB

Results (including full author names) can be downloaded in text or XML format (which is really heavy).
I’ve downloaded a small dataset (291 articles) in XML format and this is the new workflow
PubMedCentral_search_xml.knwf (35.0 KB)
The last workflow is based on the Vernalis node “European PubMed Central Advanced Search” (close but no cigar: fulltext not available AFAIK)
EuroPubMed_search_xml.knwf (36.9 KB)
Hope these workflow can help you, at least as suggestions. I think I have learned more myself than I have been able to help

1 Like