Ecosyste.ms: Awesome

An open API service indexing awesome lists of open source software.

Awesome Lists | Featured Topics | Projects

https://github.com/genepattern/dapple

Disease Association Protein-Protein Link Evaluator (Broad hosted algorithm)
https://github.com/genepattern/dapple

community genepattern-module gparc user-contri

Last synced: about 2 months ago
JSON representation

Disease Association Protein-Protein Link Evaluator (Broad hosted algorithm)

Awesome Lists containing this project

README 

        
# DAPPLE

## Background

DAPPLE stands for Disease Association Protein-Protein Link Evaluator. DAPPLE looks

for significant physical connectivity among proteins encoded for by genes in loci

associated to disease according to protein-protein interactions reported in the literature.

The hypothesis behind DAPPLE is that causal genetic variation affects a limited set of

underlying mechanisms that are detectable by protein-protein interactions. Please refer to

the DAPPLE publication for full details.

DAPPLE takes as input a list of genes, SNPs or genomic regions. See section II for a

detailed description. It will build direct and indirect interaction networks from proteins

encoded for by seed genes. It will then assess the statistical significance of a number of

network connectivity parameters as well as of the connectivity of individual proteins to

other seed proteins using a within-degree node-label permutation method. The individual

protein scores are then used to propose candidate genes in large loci. **Please note that the

most recent release uses adaptive permutation whereby networks that do not

achieve P < 0.1 (for any parameter) are stopped at 100 permutations.**

## Inputs

DAPPLE takes 4 types of input:

- A list of **genes**. These are entered as one entry per line, either in a specified file or

directly via the webpage interface. Each gene should be identified with its gene

symbol (ie Hugo) ID, such as "ATXN1". This mode should be used if the user

does not want to group genes into regions, but rather wants each gene to stand as

its own region.

- A list of **SNPs**. These are entered as one SNP per line, either in a specified file or

directly via the webpage interface. Thse SNPs must be in HapMap or 1KG,

because this is how DAPPLE defines the 'wingspan' region around a gene which

is a function of linkage disequilibrium.

- A list of **regions**. These are entered as one region per line. Each region should be

entered as 'ID chr left right' where ID is a region identifier, chr is a number from

1-23, left is the left boundary in genomic coordinates and right is the right

boundary. The entries can be space or tab delimited.

- A list of **genes with region identifiers**, or "gene-regions". These are entered as

one entry per line, either in a specified file or directly via the webpage interface.

Each entry should be defined as 'gene ID' where gene is a gene name in gene

symbol (ie Hugo) ID, such as "ATXN1", and ID refers to a region to assign the

gene to. Since DAPPLE is specifically looking for connectivity between regions -

and not within regions - the user can group genes based on how they want to

define groups.



Here are some examples:  



*Gene input*

```PANK4

HES5

TNFRSF14

MMEL1

PADI4

PTPN22

```

*SNP input*

```rs3890745

rs2240340

rs2476601

rs11586238

rs7528684

rs12746613

```

*Region input*

```rs3890745 1 2395699 2744704

rs2240340 1 17471282 17551282

rs2476601 1 113874482 114254482

rs11586238 1 117057482 117097482

rs7528684 1 155807552 156083552

rs12746613 1 159656552 159750570

```

*Gene-region input*

```PANK4 rs3890745

HES5 rs3890745

TNFRSF14 rs3890745

MMEL1 rs3890745

PADI4 rs2240340

PTPN22 rs2476601

```

*Combination input*

```rs3087243

rs6822844

region1 1 159656552 159750570

region2 1 196861967 197040967

SLC26A10

PIP4K2C

```

## Output Files

DAPPLE outputs a number of files, all of which are described here. “FILE” refers to the

keyword input by the user.



**FILE_summary**: This file contains the parameter values for the 4 network statistics

measured: (1) The number of direct connections between seen proteins from different

loci, (2) the average seed protein direct connectivity (a.k.a. direct binding degree), (3) the

average seed protein indirect connectivity (a.k.a. indirect binding degree) and (4) the

average common indiractor binding degree (the average number of seed proteins that

common interactors bind to).



**FILE_NetStats**: This file contains the permutation p-values for the 4 network statistics

described in FILE_summary (i.e., what is the probability that I see a parameter value >=

the observed value by chance?)



**FILE_SeedScores**: This file contains the individual p-values for seed proteins -

generally, the probability that by chance the seed protein would be as connected to other

seed proteins (directly or indirectly) as is observed. Please refer to the [publication's

supplementary materials](https://www.ncbi.nlm.nih.gov/pubmed/21249183) for exact details of p-value calculation. The file contains 4

columns: gene ID, region ID, uncorrected p-value, corrected p-value.



**FILE_GenesToPrioritize**: This file contains genes that achieved a corrected p-value less

than 0.05.



**FILE_CIscores**: This file contains the p-values for common interactors that describe the

probability that by chance individual common interactors would be as connected to seed

proteins as was observed.



**FILE_directConnections**: This file contains a list of the direct connections in the

network.



**FILE_plot**: If the user chose plot=true, this is the visualization of the network. Page 1

shows the direct network and pages 2-3 show the indirect network. Colors of seeds

correspond to region.



**FILE_MissingGenes**: This file is important to pay attention to. If the input is SNPs or

regions, this describes the genes in those input regions that are in the InWeb database in

contrast to those that aren't. If too many input proteins are not in the InWeb database (less

than 60% average inclusion), one should be careful about interpreting DAPPLE results.



**FILE_permuted\***: Values of permuted parameters



## FAQ

**What is DAPPLE testing?** The hypothesis behind DAPPLE is that causal genetic

variants affect common mechanisms and that these mechanisms can be inferred by

looking for physical connections between proteins encoded in disease-associated regions.

DAPPLE is therefore testing whether the networks built from seed regions - both direct

networks and indirect networks - are more connected than chance expectation. Chance

expectation is defined by the connectivity expected if connectivity were purely a function

of the binding degree of participating proteins.



**If I input a SNP, how is the region defined around that SNP?** The region is defined

using LD according to the HapMap. For a given SNP, we extend out to the region defined

by SNPs in r^2>=0.5 and then extend out to the nearest hotspots.



**For a region, how are overlapping genes defined?** The hg18 gene list was downloaded

from UCSC using Ensemble transcripts. Splice isoforms were then collapsed to define

the largest gene footprint from transcription start to transcription stop. Gene footprints

were then extended on either end to include 50kb of regulatory sequence by default,

though the user can specify a different regulatory region. Any gene footprint that overlaps

a region is included in that region. If a gene overlaps 2 regions, those regions are merged.

If the user would like to keep the regions seperate, they should input genes and explicitly

assign them to regions (option #3 on "What type of input does DAPPLE take?").



**Where does the protein-protein interaction data come from?** We use the InWeb3

databased, published by Kasper Lage in 2007 \([1](https://www.ncbi.nlm.nih.gov/pubmed/17344885),[2](https://www.ncbi.nlm.nih.gov/pubmed/27892958)\). This database contains 428,430 reported

interactions, 169,810 of which are deemed high-confidence, non-self interactions across

12,793 proteins. High-confidence is defined by a rigorously tested signal to noise

threshold as determined by comparison to well-established interactions. Briefly, InWeb

combines reported protein interactions from MINT, BIND, IntAct, KEGG annotated

protein-protein interactions (PPrel), KEGG Enzymes involved in neighboring steps

(ECrel), Reactome and others as described elsewhere in detail. All human interactions

were pooled and interactions in orthologous protein pairs passing a strict threshold for

orthology were included. Each interaction was assigned a probabilistic score based on the

neighborhood of the interaction, the scale of the experiment in which the interaction was

reported and the number of different publications in which the interaction had been sited.

## References

Rossin EJ, Lage K, Raychaudhuri S, Xavier RJ, Tartar D, IIBDGC, Cotsapas C, Daly MJ.

2011 Proteins Encoded in Genomic Regions Associated with Immune-Mediated Disease

Physically Interact and Suggest Underlying Biology. PLoS Genetics 7(1): e1001273



Lundby, A. & Rossin EJ et al. Annotation of loci from genome-wide association studies

using tissue-specific quantitative interaction proteomics. Nat. Methods 11, 868–874

(2014).

## Contact

Please email [email protected] with any questions.