Sequential Projection Pursuit Principal Component Analysis – Dealing with Missing Data Associated with New -Omics Technologies

Journal Article · Fri Mar 15 00:00:00 EDT 2013 · BioTechniques, 54(3):165-168

DOI:https://doi.org/10.2144/000113978· OSTI ID:1072899

Webb-Robertson, Bobbie-Jo M.; Matzke, Melissa M.; Metz, Thomas O.; McDermott, Jason E.; Walker, Julia; Rodland, Karin D.; Pounds, Joel G.; Waters, Katrina M.

We present a new version of sequential projection pursuit Principal Component Analysis (sppPCA) that has the capability to perform PCA on large multivariate datasets that contain non-random missing values. We demonstrate that sppPCA generates more robust and informative low-dimensional representations of the data than imputation-based approaches and improved downstream statistical analyses, such as clustering or classification. A Java program to run sppPCA is freely available at https://www.biopilot.org/docs/Software/sppPCA.

• Research Organization: Pacific Northwest National Lab. (PNNL), Richland, WA (United States)
• Sponsoring Organization: USDOE
• DOE Contract Number: AC05-76RL01830
• OSTI ID: 1072899
• Report Number(s): PNNL-SA-87092; 400412000
• Journal Information: BioTechniques, 54(3):165-168, Vol. 54, Issue 3; ISSN 0736-6205
• Country of Publication: United States
• Language: English