Biomedical subjects

Pingzhao Hu

Publications and source records attributed to Pingzhao Hu.

4 recordsLinked to original sources

PubMed2006

Computational prediction of cancer-gene function.

Most cancer genes remain functionally uncharacterized in the physiological context of disease development. High-throughput molecular profiling and interaction studies are increasingly being used to identify clusters of functionally linked gene products related to neoplastic cell processes. However, in vivo determination of cancer-gene function is laborious and inefficient, so accurately predicting cancer-gene function is a significant challenge for oncologists and computational biologists alike. How can modern computational and statistical methods be used to reliably deduce the function(s) of poorly characterized cancer genes from the newly available genomic and proteomic datasets? We explore plausible solutions to this important challenge.

Computational Biology↗

PubMed2006

Global survey of organ and organelle protein expression in mouse: combined proteomic and transcriptomic profiling.

Organs and organelles represent core biological systems in mammals, but the diversity in protein composition remains unclear. Here, we combine subcellular fractionation with exhaustive tandem mass spectrometry-based shotgun sequencing to examine the protein content of four major organellar compartments (cytosol, membranes [microsomes], mitochondria, and nuclei) in six organs (brain, heart, kidney, liver, lung, and placenta) of the laboratory mouse, Mus musculus. Using rigorous statistical filtering and machine-learning methods, the subcellular localization of 3274 of the 4768 proteins identified was determined with high confidence, including 1503 previously uncharacterized factors, while tissue selectivity was evaluated by comparison to previously reported mRNA expression patterns. This molecular compendium, fully accessible via a searchable web-browser interface, serves as a reliable reference of the expressed tissue and organelle proteomes of a leading model mammal.

Animals↗

PubMed2006

Tests for differential gene expression using weights in oligonucleotide microarray experiments.

BACKGROUND: Microarray data analysts commonly filter out genes based on a number of ad hoc criteria prior to any high-level statistical analysis. Such ad hoc approaches could lead to conflicting conclusions with no clear guidance as to which method is most likely to be reproducible. Furthermore, the number of tests performed with concomitant inflation in type I error also plagues the statistical analysis of microarray data, since the number of tested quantities in a study significantly affects the family-wise error rate. It would, therefore, be very useful to develop and adopt strategies that allow quantification of the quality of each probeset, to filter out or give little credence to low-quality or unexpressed probesets, and to incorporate these strategies into gene selection within a multiple testing framework. RESULTS: We have proposed a unified scheme for filtering and gene selection. For Affymetrix gene expression microarrays, we developed new methods for measuring the reliability of a particular probeset in a single array, and we used these to develop measures for a set of arrays. These measures are then used as weights in standard t-statistic calculations, and are incorporated into the multiple testing procedures. We demonstrated the advantages of our methods using simulated data, publicly available spiked-in data as well as data comparing normal muscle to muscle from patients with Duchenne muscular dystrophy (DMD), in which a set of truly differentially expressed genes is known. CONCLUSION: Our quality measures provide convenient ways to search for individual genes of high quality. The quality weighting strategies we proposed for testing differential gene expression have demonstrable improvement on the traditional filtering methods, the standard t-statistic and a regularized t-statistic in Affymetrix data analysis.

Data Interpretation, Statistical↗

PubMed2005

Integrative analysis of multiple gene expression profiles with quality-adjusted effect size models.

BACKGROUND: With the explosion of microarray studies, an enormous amount of data is being produced. Systematic integration of gene expression data from different sources increases statistical power of detecting differentially expressed genes and allows assessment of heterogeneity. The challenge, however, is in designing and implementing efficient analytic methodologies for combination of data generated by different research groups. RESULTS: We extended traditional effect size models to combine information from different microarray datasets by incorporating a quality measure for each gene in each study into the effect size estimation. We illustrated our method by integrating two datasets generated using different Affymetrix oligonucleotide types. Our results indicate that the proposed quality-adjusted weighting strategy for modelling inter-study variation of gene expression profiles not only increases consistency and decreases heterogeneous results between these two datasets, but also identifies many more differentially expressed genes than methods proposed previously. CONCLUSION: Data integration and synthesis is becoming increasingly important. We live in a high-throughput era where technologies constantly change leaving behind a trail of data with different forms, shapes and sizes. Statistical and computational methodologies are therefore critical for extracting the most out of these related but not identical sources of data.

Algorithms↗