JP's Notes

User Tools

Site Tools

Trace:
cls

Differences

This shows you the differences between two versions of the page.

Link to this comparison view

Both sides previous revisionPrevious revision
cls [2019/03/14 11:38] – jpcls [2019/05/23 08:49] (current) – jp
Line 12: Line 12:
  
 ^ Date/Time ^ Speaker     ^ Location ^ Title & Abstract ^ ^ Date/Time ^ Speaker     ^ Location ^ Title & Abstract ^
 +|  Fri 24th May 2019, 12:00pm-1:00pm | Dr Jean-Paul Ebejer | Faculty of ICT, Networks Lab, Level -1 block B, Room 7 | **Machine Learning in Computer-Aided Drug Discovery (Workshop)** \\ Computer-Aided Drug Design (CADD) plays an increasingly critical role in the drug-discovery process.  CADD involves the application of computer algorithms to improve pharmaceutical productivity.  These include algorithms for the identification of the biological target involved in a disease, toxicity and side-effect prediction, and searching a database for molecules which exhibit a therapeutic effect against a particular protein of interest.  The latter is known as Virtual Screening. In this workshop I will give an overview of CADD with particular emphasis on virtual screening.  We will develop a machine learning (ML) model to discriminate between actives and decoys against a protein target which plays a critical role in the life-cycle of HIV.  This interdisciplinary talk is aimed at an audience with prior Python programming experience and an interest in the application of ML models in life sciences. This talk is part of the events organised by the [[https://www.um.edu.mt/research/researchprojects/dsrg|Data-Science Research Platform]]. |
 |  Fri 15th March 2019, 2:30-3:30pm | Mr Kenneth Penza | Faculty of ICT, Room 9, Level -1, Block A | **GO term predictions in CATH: a machine learning approach** \\ Proteins are composed of amino acids chains that perform tasks such as regulation and signalling within an organism. The protein folds into a three-dimensional structure, giving it functionality. A protein function is characterised through laboratory experiments or computational methods. Physical laboratory experiments are expensive but more reliable with a low throughput. Two proteins are related if they have a common ancestor. The ancestry (homology) link is established through conserved regions in the protein sequence. Homology relationships can be utilised to transfer protein functions between related proteins. This research investigates machine learning techniques to predict protein function from protein properties and proportions from structural databases such as CATH and PFAM. This work used two Machine Learning (ML) approaches to tackle this problem. The first approach was the application of automatic feature selection using Support Vector Machines (SVM) and Random Forest (RF) techniques on species-specific datasets. The second approach applied the different species-specific datasets to Neural Networks with different hidden layer configurations. These techniques were evaluated on CAFA3 targets against the CAFA2 shared task. The RF models with the species-specific feature set performs at the same level of the best CAFA2 submission for Homo sapiens species and is superior to the best CAFA2 submission for E. coli.| |  Fri 15th March 2019, 2:30-3:30pm | Mr Kenneth Penza | Faculty of ICT, Room 9, Level -1, Block A | **GO term predictions in CATH: a machine learning approach** \\ Proteins are composed of amino acids chains that perform tasks such as regulation and signalling within an organism. The protein folds into a three-dimensional structure, giving it functionality. A protein function is characterised through laboratory experiments or computational methods. Physical laboratory experiments are expensive but more reliable with a low throughput. Two proteins are related if they have a common ancestor. The ancestry (homology) link is established through conserved regions in the protein sequence. Homology relationships can be utilised to transfer protein functions between related proteins. This research investigates machine learning techniques to predict protein function from protein properties and proportions from structural databases such as CATH and PFAM. This work used two Machine Learning (ML) approaches to tackle this problem. The first approach was the application of automatic feature selection using Support Vector Machines (SVM) and Random Forest (RF) techniques on species-specific datasets. The second approach applied the different species-specific datasets to Neural Networks with different hidden layer configurations. These techniques were evaluated on CAFA3 targets against the CAFA2 shared task. The RF models with the species-specific feature set performs at the same level of the best CAFA2 submission for Homo sapiens species and is superior to the best CAFA2 submission for E. coli.|
 |  Wed 16th January 2019, 12-1pm | Mr Nicholas Mamo | BM402 (CMMB, Biomedical Sciences Building) | **Demystifying Blockchain Technology: The Blockchain for Non-Computer Scientists** \\ If you live in Malta, then you also inhabit the self-proclaimed Blockchain Island. What does that actually mean? What is the blockchain, and how will it change the way that we think about data?  Blockchains are mentioned in the same breath as crypto-currencies, but their applicability extends beyond the world of finance. This talk breaks down the technology in non-technical terms to debunk its misconceptions and understand how it could impact healthcare. We will explain the main properties of blockchain systems, and describe requirements necessary for the correct application of blockchain technologies.  Nicholas Mamo read an undergraduate degree in AI and is presently a Research Support Officer at the Centre for Molecular Medicine and Biobanking working on the externally-funded “Connecting for Health” project.| |  Wed 16th January 2019, 12-1pm | Mr Nicholas Mamo | BM402 (CMMB, Biomedical Sciences Building) | **Demystifying Blockchain Technology: The Blockchain for Non-Computer Scientists** \\ If you live in Malta, then you also inhabit the self-proclaimed Blockchain Island. What does that actually mean? What is the blockchain, and how will it change the way that we think about data?  Blockchains are mentioned in the same breath as crypto-currencies, but their applicability extends beyond the world of finance. This talk breaks down the technology in non-technical terms to debunk its misconceptions and understand how it could impact healthcare. We will explain the main properties of blockchain systems, and describe requirements necessary for the correct application of blockchain technologies.  Nicholas Mamo read an undergraduate degree in AI and is presently a Research Support Officer at the Centre for Molecular Medicine and Biobanking working on the externally-funded “Connecting for Health” project.|
cls.1552563494.txt.gz · Last modified: 2019/03/14 11:38 by jp