Enrichment Analysis | Part 2

Loading...

来自 Icahn School of Medicine at Mount Sinai 的课程

Big Data Science with the BD2K-LINCS Data Coordination and Integration Center

12 个评分

Icahn School of Medicine at Mount Sinai

Big Data Science with the BD2K-LINCS Data Coordination and Integration Center

12 个评分

The Library of Integrative Network-based Cellular Signatures (LINCS) is an NIH Common Fund program. The idea is to perturb different types of human cells with many different types of perturbations such as: drugs and other small molecules; genetic manipulations such as knockdown or overexpression of single genes; manipulation of the extracellular microenvironment conditions, for example, growing cells on different surfaces, and more. These perturbations are applied to various types of human cells including induced pluripotent stem cells from patients, differentiated into various lineages such as neurons or cardiomyocytes. Then, to better understand the molecular networks that are affected by these perturbations, changes in level of many different variables are measured including: mRNAs, proteins, and metabolites, as well as cellular phenotypic changes such as changes in cell morphology. The BD2K-LINCS Data Coordination and Integration Center (DCIC) is commissioned to organize, analyze, visualize and integrate this data with other publicly available relevant resources. In this course we briefly introduce the DCIC and the various Centers that collect data for LINCS. We then cover metadata and how metadata is linked to ontologies. We then present data processing and normalization methods to clean and harmonize LINCS data. This follow discussions about how data is served as RESTful APIs. Most importantly, the course covers computational methods including: data clustering, gene-set enrichment analysis, interactive data visualization, and supervised learning. Finally, we introduce crowdsourcing/citizen-science projects where students can work together in teams to extract expression signatures from public databases and then query such collections of signatures against LINCS data for predicting small molecules as potential therapeutics.

从本节课中

Enrichment Analysis

This module introduces the important concept of performing gene set enrichment analyses. Enrichment analysis is the process of querying gene sets from genomics and proteomics studies against annotated gene sets collected from prior biological knowledge.

Enrichment Analysis | Part 111:59

Enrichment Analysis | Part 27:12

Enrichr Demo9:40

与讲师见面

Avi Ma’ayan, PhD
Director, Mount Sinai Center for Bioinformatics
Professor, Department of Pharmacological Sciences

So far, we've examined statistical tests for

the overrepresentation of categories in a gene set.

This leads us on to looking at

measures of the relationships between sets of genes.

So, for example genes that are found in a matter analysis to be associated

with diabetes and genes that are members pathway for example.

Now something you might see very often in the literature,

is the use of the Jaccard index.

So this is essentially a measure of the similarity between two sets.

It's a measure of how large is the overlap between the sets,

with respect to the combination of those two sets.

Here we can see, on the right an illustration of this.

Let's say we have two sets, a and b, and a is a big yellow circle and

b is a smaller set and this is the red circle.

And this, you can read this but like a Venn diagram, exactly like a Venn diagram

such that, everything that is yellow belongs only to set A,

all those genes that are a part of this region belong only to set B.

And this intersection, of the two circles are genes that belong to both sets.

So, when this terms the Jaccard index is given by the intersection, the size

of the intersection which is the number of genes that are present in both sets.

Divided by the union which is the total number of genes that we have,

total number of unique genes that we have in the two sets.

Now the Jaccard index is a number that varies between zero and one.

The value is 1 when the sets are identical, and

the value is 0 when there is no overlap between the two sets.

The analysis of the enrichment of gene sets is something that perhaps,

has developed from over representation analysis.

And it is a little bit more of a sophisticated approach.

It has a similar aim in that it attempts

to find relevant gene sets.

But it does it in a.

Perhaps the most sophisticated way, such that we don't draw hard thresholds here.

And so we don't, for example, just take the top most deferentially

expressed genes and use those to determine the gene sets.

We actually use the whole gene expression profile in a weighted fashion.

So, here, I can give a brief overview,

of the aims and the advantages of this approach.

So, a microarray profile measures

the expression levels of a large number of genes.

So, something on the scale of the whole genome.

And a typical biological question is to try to

understand the difference in expression between two sets of samples.

So these might be a diseased and non-diseased or control samples and

samples that have undergone some kind of perturbation.

Traditionally, in the first instance, we might look for

individual genes which are differentially expressed between the two classes.

And then use that as the basis of further analysis,

perhaps overrepresentation analysis.

There are two problems, potential problems with this.

The first is the biological data,

generally gene expression data, can be noisy.

There are lots of sources of variation, and

no single gene, it may be the case that there are no genes

stand out above the noise as being significantly differentially expressed.

Alternatively, many, many genes might be identified as deferentially

expressed but it's difficult then to interpret

the list of genes to infer

a unified biological theme for these genes.

So, there was a problem of biology interpretation.

The analysis in the enrichment's of gene sets, methods such as GSEA.

Instead of looking for individual genes that are deferentially expressed.

Look for whole collections of genes that are collectively

differentially expressed and this has two main advantages.

The first is that there is more statistical power here.

So, a set of genes is more likely to stand out above the noise than individual genes.

We expect a larger signal to noise ratio, in other words.

Secondly, if the gene sets are chosen

carefully such that they all belong to

a particular biological theme,

then the biological interpretation

is built into the analysis or

it's almost automatic.

[MUSIC]

探索我们的目录

免费加入并获得个性化推荐、更新和优惠。

Coursera 致力于普及全世界最好的教育，它与全球一流大学和机构合作提供在线课程。

© 2018 Coursera Inc. 保留所有权利。

通过 App Store 下载

通过 Google Play 获取