Bioinformatics Seminar
Time:
Venue: Na
16 April 2019
NaMethods for imputation
Soroor Hediyeh-zadehWEHI Bioinformatics
differential expression and multi-Omics integration of label-free proteomics data;High-throughput mass spectrometry (MS) data are affected by a variety of known and unknown systematic biases as well as high proportions of missing values. Systematic bias is inherent in MS-based data due to complex biological ; experimental and technical processing. Bias ; which may be loosely defined as any non-biological signal ; may occur due to many factors ; including variations in sample processing conditions ; instrument calibrations ; etc. ; and manifests in MS measurement accuracy ; retention times and peak intensities.
We address the current challenges in the analysis of label-free proteomics profiles ; propose a new technique for imputation of missing values ; and explain how the existing tools for the analysis of microarray data based on linear models and empirical Bayes moderated t-statistics can be leveraged to improve the differential expression analyses at the peptide or protein level. We then motivate methods for the integration of shotgun proteomics data with other High-throughput data ; namely RNA-seq ; by mapping the datasets into a common lower-dimensional embedding ; where they share similarities. The embedding is then used to infer the biological associations between the features (i.e. proteins ; peptides ; genes ; transcripts) of the datasets.