Javascript is disabled

Data-analysis Workshop                         
November 12^th 2017

Organized by Swedish Proteomics Society, BioMS and SciLifeLab
Moderator: Janne Lehtiö, Karolinska Institutet, Stockholm, Sweden

11.00-12.30 What drives error rates in mass spectrometry-based proteomics?
Lukas Käll, Royal Institute of Technology, Stockholm, Sweden

In this lecture, we will discuss a couple of points related to statistics and machine learning for improving identification and quantification in shotgun proteomics. First, we will discuss different measures of error rates of identifications and quantifications, such as false discovery rates, and the factors of database searching that influence such measures. In particular, we will discuss the relationships between score functions, search space and error rates. Secondly, we will discuss search space reduction techniques and machine learning methods that aim to improve yield and reliability of experiments, such as our semi-supervised machine learning method, Percolator, and our unsupervised fragment spectra-clustering method, MaRaCluster. 

12.30- 13.30 Lunch

13.30-15.00 Tools for quantitative analysis of proteome
Yafeng Zhu and Henrik Johansson, Karolinska Institutet, Stockholm, Sweden

Mass spectrometry based proteomic has become widely used for protein quantification due to improved proteome coverage in recent years. The development of quantitative analysis tools has been mostly solving two major problems: protein abundance summarization and statistical test to detect protein changes. Here we discuss pros and cons of a few selected tools and present a new tool, DEqMS, recently developed in our group for testing differential protein expression while accounting variance heteroscedasticity in proteomics data.

15.00-16.30 Cancer Proteogenomics
David Fenyö, Institute for Systems Genetics, NYU School of Medicine, USA

Advances in sequencing technologies have revealed large heterogeneity on the genome and transcriptome level in tumors. However, it has often been difficult pinpoint which of the changes are important drivers of tumor growth. Proteomic technologies have also improved rapidly, and although they are still have a way to go before providing the same dynamic range and sensitivity as sequencing technologies, they provide rich complementary information. The combined application of proteomics and genomics to the understanding of tumor biology has the potential of driving innovative diagnostics and new treatments for cancer. I will discuss the integration of data from breast and ovarian tumors analyzed within The Cancer Genome Atlas (TCGA) and the Clinical Proteomics Tumor Analysis Consortium (CPTAC).

Venue (Workshop)

SciLifeLab, Tomtebodaväg 23a, Solna

Fee: 500 sek exkl VAT

Registration
Registration is made when you register for the SPS 2017conference.
Click here to register

Already registered for the SPS conference and want to join the Workshop?
Send an e-mail to lotta@mkon.se and we will add the workshop to your registration.