Abstract
Original language | English |
---|---|
Title of host publication | Discovery Science |
Publisher | Springer |
Pages | 299-307 |
ISBN (Electronic) | 978-3-319-24282-8 |
ISBN (Print) | 978-3-319-24281-1 |
DOIs | |
Publication status | Published - 2015 |
MoE publication type | A4 Article in a conference publication |
Event | 18th International Conference on Discovery Science, DS 2015 - Banff, Canada Duration: 4 Oct 2015 → 6 Oct 2015 Conference number: 18 |
Publication series
Series | Lecture Notes in Computer Science |
---|---|
Volume | 9356 |
Conference
Conference | 18th International Conference on Discovery Science, DS 2015 |
---|---|
Abbreviated title | DS 2015 |
Country | Canada |
City | Banff |
Period | 4/10/15 → 6/10/15 |
Fingerprint
Keywords
- Canonical correlation
- Kernel methods
- Sparsity
- Deep biosphere
Cite this
}
Canonical correlation methods for exploring microbe-environment interactions in deep subsurface. / Uurtio, Viivi; Bomberg, Malin; Nybo, Kristian; Itävaara, Merja; Rousu, Juho.
Discovery Science. Springer, 2015. p. 299-307 (Lecture Notes in Computer Science, Vol. 9356).Research output: Chapter in Book/Report/Conference proceeding › Conference article in proceedings › Scientific › peer-review
TY - GEN
T1 - Canonical correlation methods for exploring microbe-environment interactions in deep subsurface
AU - Uurtio, Viivi
AU - Bomberg, Malin
AU - Nybo, Kristian
AU - Itävaara, Merja
AU - Rousu, Juho
PY - 2015
Y1 - 2015
N2 - In this study, we apply non-linear kernelized canonical correlation analysis (KCCA) as well as primal-dual sparse canonical correlation analysis (SCCA) to the discovery of correlations between sulphate reducing bacterial taxa and their geochemical environment in the deep biosphere. For visualization of canonical patterns, we demonstrate the applicability of the correlation plot technique on kernelized data. Finally, we provide an extension to the visual analysis by clustergrams. The presented framework and visualization tools enabled extraction of latent canonical correlation patterns between the salinity of the groundwater and the bacterial taxonomic orders Desulfobacterales, Desulfovibrionales and Clostridiales.
AB - In this study, we apply non-linear kernelized canonical correlation analysis (KCCA) as well as primal-dual sparse canonical correlation analysis (SCCA) to the discovery of correlations between sulphate reducing bacterial taxa and their geochemical environment in the deep biosphere. For visualization of canonical patterns, we demonstrate the applicability of the correlation plot technique on kernelized data. Finally, we provide an extension to the visual analysis by clustergrams. The presented framework and visualization tools enabled extraction of latent canonical correlation patterns between the salinity of the groundwater and the bacterial taxonomic orders Desulfobacterales, Desulfovibrionales and Clostridiales.
KW - Canonical correlation
KW - Kernel methods
KW - Sparsity
KW - Deep biosphere
U2 - 10.1007/978-3-319-24282-8_25
DO - 10.1007/978-3-319-24282-8_25
M3 - Conference article in proceedings
SN - 978-3-319-24281-1
T3 - Lecture Notes in Computer Science
SP - 299
EP - 307
BT - Discovery Science
PB - Springer
ER -