Comments (4)
Sorry for the inconvenience. You can find the description as follows. I omit some elements which are not frequently used.
- B
latent variable matrix with dimension k-by-samples. You can correlate this matrix with the phenotype of interest
- Z
loading matrix with dimension genes-by-k. You can check which genes are top ranked with respect to each LV by ordering the loading values (decreasing order) in a column-wise manner
- U
coefficient matrix with dimension pathways-by-k. This encodes the combination of pathways for each LV.
- C
The exact pathway matrix used in the decomposition which may be different from the input
- withPrior
An index vector which indicates the list of LVs that are associated with pathways
- Uauc
see Up
- Up
Uauc and Up are AUC and p.val matrices with the same dimension as U. They are calculated based on cross-validation to test how reliable pathways are associated with LVs. All entries with non-zero AUC are collected to formulate summary
- summary
This is based on Uauc and Up which records pathway-LV pairs that are supported by cross-validation test. For a given LV, if you find the associated pathways are interesting, you may want to double check the top ranked genes for this LV by looking at Z
from plier.
Thanks, I am wondering about how I should interpret 'summary'. It looks like not all of the pathways from Up and Uauc are represented in 'summary', it seems to be because they have zero AUC. Does the cross validation test you describe refer to the "Gene-holdout cross-validation" section in your paper? if so,
- If a pathway is present in Uauc but not 'summary', should we interpret this as 'the pathway is not well represented among the LVs' and/or 'the pathway is not found to be highly expressed in this dataset'?
- Since I have 386 LVs in Uauc, is it correct to say that this was the number of LVs found after convergence, and is preserved throughout cross validation? However number of unique LV indexes in 'summary' is 262. Is it useful to perhaps try functional enrichment on LVs that were not assigned to any pathway?
- Do you think it is necessary to do any kind of multiple testing correction on the p-values?
from plier.
Yes, it refers to the "Gene-holdout cross-validation" section.
- It means there are no LVs that are associated with this pathway based on cross validation test. You may double check this by performing enrichment analysis, and/or check top ranked genes. (check the next bullet)
- Cross validation was conducted after the decomposition finished, so these 386 LVs were just the result of convergence. There is a parameter
frac=0.7
which controls the fraction of LVs that have at least 1 prior information association. That's why the number is 262. Performing pathway/function enrichment on loading (Z
in the output) is recommended as a double check on all LVs. Some LVs don't have pathway association based on cross-validation, but they may be a valid pathway surrogate by looking at top-ranked genes. - The FDR values are already included in the
summary
as one column.
from plier.
Okay I see that there are FDR assigned to each pathway. Thanks very much Wayne.
from plier.
Related Issues (16)
- Q-value error HOT 3
- rseed argument gives "error in Z %*% B : non-conformable arguments" HOT 2
- Adding Release Tag HOT 6
- Issue while running Tutorial HOT 8
- Prior data HOT 6
- Recommendation on the size of prior data? HOT 3
- Several questions about PLIER HOT 4
- Does these errors occur due to the small size of my dataset? HOT 2
- making a new pathways matrix HOT 14
- Problem with installing Rd object HOT 1
- Two question regarding interpretation of the pathways having significant AUC for a specific LV HOT 1
- Should the input gene expression data be log scale or not? HOT 2
- try to apply PLIER to seurat object: Error in qr.default(Y, complete = FALSE) : NA/NaN/Inf in foreign function call (arg 1) HOT 1
- Can the data of mRNA of human tissue be use this method? HOT 8
- PLIER::num.pc error HOT 4
Recommend Projects
-
React
A declarative, efficient, and flexible JavaScript library for building user interfaces.
-
Vue.js
🖖 Vue.js is a progressive, incrementally-adoptable JavaScript framework for building UI on the web.
-
Typescript
TypeScript is a superset of JavaScript that compiles to clean JavaScript output.
-
TensorFlow
An Open Source Machine Learning Framework for Everyone
-
Django
The Web framework for perfectionists with deadlines.
-
Laravel
A PHP framework for web artisans
-
D3
Bring data to life with SVG, Canvas and HTML. 📊📈🎉
-
Recommend Topics
-
javascript
JavaScript (JS) is a lightweight interpreted programming language with first-class functions.
-
web
Some thing interesting about web. New door for the world.
-
server
A server is a program made to process requests and deliver data to clients.
-
Machine learning
Machine learning is a way of modeling and interpreting data that allows a piece of software to respond intelligently.
-
Visualization
Some thing interesting about visualization, use data art
-
Game
Some thing interesting about game, make everyone happy.
Recommend Org
-
Facebook
We are working to build community through open source technology. NB: members must have two-factor auth.
-
Microsoft
Open source projects and samples from Microsoft.
-
Google
Google ❤️ Open Source for everyone.
-
Alibaba
Alibaba Open Source for everyone
-
D3
Data-Driven Documents codes.
-
Tencent
China tencent open source team.
from plier.