Polska Infrastruktura Informatycznego Wspomagania Nauki w Europejskiej Przestrzeni

Polska Infrastruktura Informatycznego Wspomagania Nauki w Europejskiej Przestrzeni Badawczej The PL-Grid Virtual Laboratory in the Life Sciences Domain Maciej Malawski, Eryk Ciepiela, Tomasz Gubała, Piotr Nowakowski, Daniel Harężlak, Marek Kasztelnik, Joanna Kocot, Tomasz Bartyński, and Marian Bubak Institute of Computer Science AGH ACC Cyfronet AGH

Outline u Motivation – complex scientific applications on modern u u computing infrastructures In-silico experiments and Virtual Laboratory Grid. Space 2 as a solution ─ Architecture ─ Working with Grid. Space Examples of applications ─ Computational chemistry ─ Bioinformatics Conclusions 2

Motivation u Complex scientific applications on modern computing infrastructures ─ Clusters, Grids, Clouds u Diverse software packages ─ Applications (Gaussian, NAMD, …) ─ Web Services ─ Scripts: Perl, Python, Ruby u Different users ─ Chemists, biologists ─ Programmers ─ End users u Various data types ─ Files, databases, URLs u Exploratory programming ─ Unstructured, dynamic, prototyping u Collaboration ─ Teams, communities 3

Experiment u Experiment (in-silico)- a process that combines together data with a set of activities (programs, services) that act on that data in order to produce experiment results ─ ─ u u u Experiment plan – a specific type of software Experiment run – a specific execution of the experiment Complex workflow going beyond manual simple and repeatable execution of installed programs Combines steps realized on a range of software environments, platforms, tools, languages etc. Developed, shared and reused collaboratively amongst ad-hoc researching teams Composed of collaboratively owned libraries and services used (called gems) and experiment parts (called snippets) Virtual Laboratory – environment for development, execution and sharing of experiments 4

Working with Grid. Space 2 u Easy access using Web browser u Experiment Workbench ─ ─ Constructing experiment plans from code snippets Interactively run experiments u Experiment Execution Environment ─ ─ Multiple interpreters Access to libraries, programs and services (gems) u Access to computing infrastructure ─ Cluster, grid, cloud 5

Experiment Workbench 6

Binding sites in proteins u Comparison of Services for Predicting Ligand Binding Sites ─ ─ Multiple services available on the Web Conversions between data formats Visualization scripts (Jmol, Gnuplot) Single access based on experiments developed in Virtual Laboratory u Calculation of hydrophobicity profiles ─ Multiple scales, parameters, input data ─ Computed using PL-Grid resources – easy access to Zeus cluster at Cyfronet ─ Management of experiment results: ~ 1 Million output files ─ Using semantic integration framework for metadata management Collaboration with Department of Bioinformatics and Telemedicine, Jagiellonian University, Prof. Irena Roterman-Konieczna, Katarzyna Prymula 7

Analysis of water solutions of aminoacids u Involving multiple steps realized with many tools, languages and libraries used for ─ ─ ─ Packmol – molecular dynamics simulations of packing molecules in a defined regions of space Jmol – visualization of solution Gaussian – computing a spectrum of the solution Python/CCLIB – extracting spectrum info jq. Plot – displaying plot Collaboration with computational chemists of ACC Cyfronet AGH and Department of Chemistry, Jagiellonian University, Dr. Mariusz Sterzel, Klemens Noga 8

Conclusions u Complex scientific applications need dedicated tools and approaches. u In-silico experiments are supported by Virtual Laboratory powered by Grid. Space 2 technology. u Applications: ─ ─ ─ Bioinformatics Computational chemistry More are welcome! u Virtual laboratory is open for PL-Grid users. 9

References u http: //wl. plgrid. pl – open the Virtual Laboratory in your browser u http: //gs 2. cyfronet. pl – learn more about Grid. Space 2 technology u http: //virolab. cyfronet. pl – see our earlier achievements u http: //www. plgrid. pl – become a user of PL-Grid 1 0