Sign inSign up
Metabolite identification and molecular fingerprint prediction via machine learning thumbnail
Pause
Mute
Subtitles
Playback speed
0.25
0.5
0.75
1
1.25
1.5
1.75
2
Full screen

Metabolite identification and molecular fingerprint prediction via machine learning

Published on Oct 23, 20122808 Views

Markus Heinonen

**Motivation:** Metabolite identification from tandem mass spectra is an important problem in metabolomics, underpinning subsequent metabolic modelling and network analysis. Yet, currently this task

MLSB 2012 - Basel

Related categories

Molecular BiologyKernel Methods

Chapter list

Metabolite identification and molecular fingerprint prediction via machine learning00:00
Outline00:10
Summary00:23
Contents: Motivation00:55
Metabolomics bottlenecks00:56
Metabolite identification01:54
Tandem mass spectrometry (MS/MS)03:47
Current metabolite identification methods05:17
Contents: Kernel framework06:51
Machine learning problem06:53
Overview of the framework08:10
Fingerprints08:39
Mass spectral kernels09:34
Integral mass kernel11:11
Spectral density model11:55
High resolution probability product kernel13:17
Fingerprints into metabolites14:47
Poisson-Binomial model15:52
Contents: Experiments17:29
Experiments17:33
Fingerprint prediction accuracy19:09
Fingerprint prediction accuracy cont.20:56
Ranks22:20
Comparison to MetFrag23:21
Conclusions24:28
Thank you24:59
Individual fingerprint prediction accuracies26:28