Affiliations: DIPAV Dipartimento di Patologia Animale, Igiene e
Sanità Pubblica Veterinaria, Università degli Studi di Milano,
Italy | CRS4 Bioinformatics Laboratory, Parco Technologico
Sardegna Ricerche, 09010 Pula (CA), Italy | Proteotech s.r.l., Parco Technologico Sardegna
Ricerche, 09010 Pula (CA), Italy | Laboratorio di Proteomica, Istituto Sperimentale
Italiano L. Spallanzani, Milano, Italy
Abstract: Typical high-abundant proteins, including albumin, IgG, IgA and
others, are the target of depletion methods usually applied to two-dimensional
electrophoresis (2DE) of human biological fluids like serum and plasma.
Detection of low-abundant proteins is of interest with regard to biomarkers for
disease when being studied by 2DE or liquid chromatography-mass spectrometry
(LC/MS). After depletion of very abundant proteins, serum samples consist of an
enriched pool of low-abundant proteins that can be further studied without
significant interferences, thus allowing for a full identification of the low
abundant proteins, whose spots become now more visible. We have employed wavelet-based techniques and their derived
denoisers to explore 2DE from disease-control human samples. We have pursued
the goal of mimicking in silico the spot detection performance experimentally
obtained by depletion methods, thus hoping to read through the critical
high-abundant protein regions. Our results suggest that an efficient and
effective computational tool has been added to other ones performing 2DE image
analysis, such as decomposition and segmentation, but with the advantage of
being specifically targeted to the depletion task.
