Graphical Models for HIV Vaccine Design
author:
David Heckerman,
Microsoft Research
Description
I will discuss two applications of graphical models to HIV vaccine design. The first helps determine how strongly our immune system fights HIV. The second helps identify which parts of HIV can be successfully attacked by our immune system. I will also discuss how these applications have exposed a weakness in the process of learning graphical models from data---namely, the inability to quantify how many arcs in a learned graphical model are spurious. I will offer a solution based on the False Discovery Rate.
You might be experiencing some problems with Your Video player.
| Slides | |
| 0:00 | Graphical models for HIV vaccine design |
| 2:55 | The need for an HIV vaccine |
| 3:35 | Overview |
| 3:57 | HIV life cycle |
| 4:45 | Two arms of adaptive immune response |
| 5:23 | Central question in vaccine design: Can our immune system stop HIV? |
| 6:09 | Cellular arm details - 1 |
| 6:43 | Cellular arm details - 2 |
| 6:54 | Cellular arm details - 3 |
| 7:02 | Cellular arm details - 4 |
| 7:14 | How effective is this mechanism on HIV? - 1 |
| 7:28 | How effective is this mechanism on HIV? - 2 |
| 8:24 | HIV mutates rapidly |
| 9:26 | Rapid mutation + selection pressure = detectible footprint |
| 10:10 | First approach - 1 |
| 10:54 | First approach - 2 |
| 12:05 | This approach ignores the phylogeny |
| 12:41 | Problem: Simple method ignores the phylogenetic structure of the data - 1 |
| 13:11 | Problem: Simple method ignores the phylogenetic structure of the data - 2 |
| 13:56 | Problem: Simple method ignores the phylogenetic structure of the data - 3 |
| 14:38 | A graphical model approach |
| 15:44 | Phylogenetic tree |
| 16:37 | Model 1: Explained by phylogeny alone |
| 18:45 | Model 2: Explained by phylogeny and HLA |
| 19:45 | Many possible associations to investigate |
| 20:15 | What biologists don’t want |
| 21:10 | What biologists want - 1 |
| 21:36 | False discovery rate (FDR) - 1 |
| 23:01 | False discovery rate (FDR) - 2 |
| 25:21 | Creating null data via permutation |
| 25:59 | FDR applied to synthetic data - 1 |
| 26:34 | FDR applied to synthetic data - 2 |
| 26:48 | FDR applied to real data - 1 |
| 28:04 | How many true associations are missing? |
| 28:25 | Associations missed |
| 28:43 | FDR applied to real data - 2 |
| 30:28 | Other insights |
| 32:33 | Can we find the epitopes? |
| 33:43 | Find the epitope HLA alleles |
| 34:17 | Example |
| 34:36 | Complications… |
| 35:32 | Graphical models to the rescue |
| 36:43 | What biologists want - 2 |
| 37:13 | FDR applied to arcs in a DAG model |
| 39:09 | Results on synthetic data |
| 39:53 | Results on real data |
| 41:22 | Conclusions |
| 42:17 | - Questions |
| 50:04 | - Questions |
Lecture rating
| People found this lecture: | ||
| Worth seeing | ||
| because it is: | ||
| Valuable and informative | ||
| Well presented | ||
| Easily understandable | ||
| Acceptably recorded | ||
| You need to login to cast your vote. | ||
Report a problem or upload files
If you have found a problem with this lecture or would like to send us extra material, articles, exercises, etc., please use our ticket system to describe your request and upload the data.Enter your e-mail into the 'Cc' field, and we will keep you updated with your request's status.
Related content
Visitors who watched this lecture also watched...
3 comments
SEE ALSO:
Download slides:
icml07_heckerman_gmhi_01.pdf (1.2 MB)
Download slides:
icml07_heckerman_gmhi_01.pptx (2.5 MB)
Launch in a standalone WM PlayerSwitch to Windows Media Player
View slides
Streaming Video Help
Windows Media Player Firefox Plugin - Download
Link this page
Would you like to put a link to this lecture on your homepage?Go ahead! Copy the HTML snippet !
