p3 training and_life_as_a_postdoc_(felix_klein)

Bioinformatics Career Day24 May 2012

Felix Klein

Felix Klein24.05.20122

Background

• physics diploma, University of Heidelberg

• diploma thesis in radiation dosimetry

at DKFZ

• measurements at HIT


Why bioinformatics?

• interdisciplinary

• programmed in R

• worked on data analysis


Progress in science is driven by technology


Chromatin loops


Investigation of chromatin 3D structure

• role of chromatin 3D structure in gene regulation

• 4C to investigate detailed interactions of

cis-regulatory modules (CRMs)

• global chromatin interactome using HiC


Investigation of chromatin 3D structure

8

Automated analysis of microscopy based RNAi screens

Imaging Segmentation

Calibrated image Segmentation mask

g.x g.y g.s g.p g.pdm [1,] 123.1391 3.288660 194 67 9.241719 [2,] 206.7460 9.442248 961 153 20.513190 [3,] 502.9589 7.616438 219 60 8.286918 [4,] 20.1919 22.358418 1568 157 22.219461 [5,] 344.7959 45.501992 2259 233 35.158966 [6,] 188.2611 50.451863 2711 249 28.732680 [7,] 269.7996 46.404036 2131 180 26.419631 [8,] 106.6127 58.364243 1348 143 21.662879 [9,] 218.5582 77.299007 1913 215 25.724580[10,] 19.1766 81.840147 1908 209 26.303760[11,] 6.3558 62.017647 340 68 10.314127[12,] 58.9873 86.034128 2139 214 27.463158[13,] 245.1087 94.387405 1048 123 18.280901[14,] 411.2741 109.198678 2572 225 28.660816[15,] 167.8151 107.966014 1942 160 24.671533[16,] 281.7084 121.609892 2871 209 31.577270

Features extraction

Object featuresObjects labels

ClassificationSummary

g.x g.y g.s g.p g.pd

123.1391 3.288660 194 67 9.241719

Phenotypic profile

Source image

aft apt neg

int pos


What was important for me?

• bioinformatics group with members of diverse backgrounds

• PI who successfully trained bioinformaticians

• well established group in bioinformatics


What might be interesting for you

• turn data into biology

• interaction with people from biology groups

• communication skills !!!

• workload divides mainly into:• programming (50 %)• reports, meetings, email


Acknowledgements

Wolfgang Huber

Simon AndersJoseph BarryBernd FischerJulian GehringAleksandra PekowskaPaul Theodor PylAlejandro ReyesMaria Secrier

Collaborators:

Michael BoutrosChristian Volz

Eileen FurlongYad Ghavi Helm

Data production rates

LHC: 1.8 GB / s at peak capacity (i.e. actively conducting a primary aspect of the LHC’s four main experiments: ATLAS, ALICE, CMS, and LHCb).

These experiments will take roughly a decade to complete, and each of them is expected to produce over a 1 PB per year of data.

One Illumina HiSeq: up to 600 Gb/run , i.e. ~600 GB/10 days = 18 TB/year (not including derived data e.g. BAM)

One Digital Embryo (2008): 3.5 TB (2048 x 2048 x 370 x 1226)

EMBL-EBI: in 9/2011, data storage capacity was 14 PB

p3 training and_life_as_a_postdoc_(felix_klein)

Technology

felix klein

investigation of chromatin

chromatin loops5

derived data

data analysis324

d structure role of

data storage capacity

data production rateslhc