Using similar data retrieval methods described above, we built a corpus of 48 news articles related to corruption, manually retrieved and annotated metaphor tokens in all the articles9.
After collecting the raw retrievals in SUMSearch and Google Scholar, we defined a reference standard by manually reviewing all retrievals (links) from the retrieval list of each search engine.
The remaining markers were stained manually with antigen retrieval in a pressure cooker followed by primary incubation overnight at 4°C.
This paper also reports on our impact evaluations on the retrieval effectiveness of manually assigned descriptors (MeSH or Medical Subject Headings), showing that by including these terms retrieval performance can improve from 2.4% to 13.5%, depending on the underling IR model.
Antigen retrieval was done manually for all antibodies except Ang-2, by placing the specimens in 0.01 M citrate buffer at pH 6.0 and exposed to microwave heating of 20 minutes at 450 W. All antibodies were incubated overnight at 4°C except for VEGF-A where the primary antibody was incubated for 30 min in room temperature.
Only demographic data of patients were stored in the hospital database to enable retrieval of files manually based on patient codes.
All annotated terms in the training corpus were automatically extracted and integrated into a seed biomarker retrieval dictionary by manually assigning them to the six main classes.
Sections of FFPE murine kidney tissue (N = 3) were manually microdissected, treated with antigen retrieval, and incubated overnight with PNGase F (two of three sections).
Antigen retrieval was performed manually with 10 m M citrate buffer (pH 6.0) in a pressure cooker at 120 °C for 10 min (PSEN2 and ADAM17) or in TEC buffer (Tris-EDTA-Citrate pH 7.8) using the DakoCytomation Pascal pressure chamber (DakoCytomation; Hamburg, Germany) at 125 °C for 1 min (MLH1, MSH2, and MSH6), respectively.
This highlights the importance of using text mining for a comprehensive retrieval of information, complementing manually annotated data, which may potentially miss over 1,700 TC-related genes (84% of 2,071).
