CLAIM-MS - CLAIM Multi Source, an expanded version of CLAIM
Authors: Marek Blazewicz1,2, Giovanni Felici3, Aleksandra Swiercz1,4, Daniele
Santoni3, Marcin Jaroszewski1, Agnieszka Zmienko1,4, Marta Kasprzak1,4
CLAIM-MS is a method for finding functionally related genes. The novelty
of this proposition is in its flexibility, as the method integrates
information from many input data sources of different types. We
successfully validated it on gene expression data produced by
different technologies (microarray, RNA-seq) and experiment setups
(case-control or multi-class, single-time-point or time-series), on
protein-protein interaction networks and Gene Ontology annotations. For
each dataset, a gene-gene distance metric needs to be derived in
accordance with its nature and the experiment setup.
This approach expands our previous work with, among others:
- the ability to handle more than two data sources at once;
- a new robustly converging clustering algorithm (a neural gas method);
- a more efficient clique detection algorithm;
- deep analysis of underlying distance matrices, which allow tuning up
the evaluation of gene clusters with respect to a particular biological
dataset; this procedure significantly improves the overall quality of
The instruction on how to run the application can be found at: README
The research was supported by grant No. 2012/05/B/ST6/03026 from
the National Science Centre, Poland. A publication presenting both the
method and the results is in preparation.
1 Institute of Computing Science,Poznan University of Technology, Poznan, Poland.
2 Poznan Supercomputing and Networking Center, Poznan, Poland.
3 Institute for Systems Analysis and Computer Science “Antonio Ruberti”, National Research Council of Italy, Rome, Italy.
4 Institute of Bioorganic Chemistry, Polish Academy of Sciences, Poznan, Poland.