Ingrid Måge, Tormod Næs and Kristian Hovde Liland attended the Nordic Arctic workshop in Groningen, in the Netherlands, together with Ulf Geir Indahl from NMBU.
There were several good talks and fruitful discussions by the attendees from Norway, Denmark, the Netherlands, Italy and Spain, spanning a variety of methods and subjects in data analysis.
Kristian Hovde Liland talks about extensions of the ROSA method for multiblock analysis.
Ulf Geir Indahl talks about Tikhonov Regularisation and fast leave-one-out cross-validation.
The 2016 AgroStat symposium was held at Nestlé Research Center in Lausanne, Switzerland on March 21-24. It was professionally organized and had a lot of high quality scientific contributions in chemometrics, sensometrics, big data and risk & process. Most of the presentations and all posters were in English, while a few exceptions were French. The first day was reserved for a variety of workshops. There were around 120 participants, mostly French and Swiss, but also from Canada, Great Britain, Scandinavia, Portugal, Italy, USA, Germany and some other countries.
From Nofima Kristian Hovde Liland contributed with a poster describing the R package MatrixCorrelation and the newly developed Similarity of Matrices Index.
Presentations, posters and short papers are available from:
The HotPLS toolbox for MATLAB is now available. It contains a set of functions for performing classification in a fixed hierarchy. For more information see:
Kristian Hovde Liland has authored a paper on classification in a fixed hierarchy using PLS-based methods together with Achim Kohler and Volha Shapaval. The paper is titled: “Hot PLS—a framework for hierarchically ordered taxonomic classification by partial least squares” and was recently published in the journal Chemometrics and Intelligent Laboratory Systems.
- Classification in a fixed hierarchy
- Utilization of replicate measurements for improved robustness through majority voting
- Automatic model building and complexity estimation from taxonomic information
- Detection of outliers and samples representing new classes absent in calibration
A novel framework for classification by partial least squares in a fixed hierarchy is presented. The hierarchical approach ensures flexible local modelling with varying complexity. It results in an intuitive classification path from the highest taxonomic levels down to species and beyond. Results are presented as phylogenetic trees with local diagnostic information to gain maximum information about the classification and help the researcher to focus on interesting phenomena.
Information on sample replicates is included in the classification to increase performance and avoid misclassifications due to low quality measurements. Detection of samples coming from previously unobserved classes is enabled by estimating cut-off distances from the calibration data classes. To further increase flexibility and improve customization the canonical powered partial least squares algorithm is used for modelling and classification together with linear discriminant analysis. This opens up for additional sample response information and forced sharpening of focus on important variables. The presented framework is not limited to biological taxonomy, but was first developed for this purpose.
- Partial least squares
- Fixed hierarchy
- Local modelling
- Replicate measurements
Kristian Hovde Liland, Achim Kohler, Volha Shapaval, Hot PLS—a framework for hierarchically ordered taxonomic classification by partial least squares. Chemometrics and Intelligent Laboratory Systems, Volume 138, 15 November 2014, Pages 41–47.
The Open EMSC toolbox is now available. For more information see:
Kristian Hovde Liland has co-authored a paper on variable selection in aligned 16S rRNA sequences together with Hilde Vinje, Trygve Almøy and Lars Snipen. The paper is titled: “A systematic search for discriminating sites in the 16S ribosomal RNA gene” and was recently published in the journal Microbial Informatics and Experimentation. This was the last paper to ever be accepted in the journal. Whether this was due to the journal editors feeling that the peak of their careers had been met with this paper and therefore decided to close the journal has not been confirmed.
The 16S rRNA is by far the most common genomic marker used for prokaryotic classification, and has been used extensively in metagenomic studies over recent years. Along the 16S gene there are regions with more or less variation across the kingdom of bacteria. Nine variable regions have been identified, flanked by more conserved parts of the sequence. It has been stated that the discriminatory power of the 16S marker lies in these variable regions. In the present study we wanted to examine this more closely, and used a supervised learning method to search systematically for sites that contribute to correct classification at either the phylum or genus level.
- Partial least squares
- Selectivity ratio
- Sequence alignment
Hilde Vinje, Trygve Almøy, Kristian Hovde Liland, Lars Snipen,
A systematic search for discriminating sites in the 16S ribosomal RNA gene
Microbial Informatics and Experimentation 4(2) (2014).
The paper “Distribution based truncation for variable selection in subspace methods for multivariate regression” has been accepted for publication in Chemometrics and Intelligent Laboratory Systems. It is authored by Kristian Hovde Liland and Harald Martens of our group together with former co-worker Martin Høy and Solve Sæbø from the Norwegian University of Life Sciences.