Introduction to Weighted Gene Co-expression Network Analysis (WGCNA) | Bioinformatics 101

Weighted Gene Co-expression Network Analysis (WGCNA) is a commonly used unsupervised method to cluster genes based on their expression profiles. In this video I go over the idea behind WGCNA and provide a high-level overview of various steps that go into this analysis. In addition, I talk about good practices when performing WGCNA. I hope you find this video helpful! Leave your thoughts in the comment section below!
Chapters:
0:00 Intro
1:01 Studying systemic response
3:04 Basic idea behind WGCNA
4:00 Examples of functional association among co-expressed genes
6:27 Terminologies
7:44 WGCNA workflow steps (high level overview)
9:19 WGCNA Applications
9:54 Good practices
You can show your support and encouragement by buying me a coffee:
www.buymeacoffee.com/bioinfor...
To get in touch:
Website: bioinformagician.org/
Github: github.com/kpatel427
Email: [e-mail]
#bioinformagician #bioinformatics #wgcna #coexpressionnetworks #geneexpression #sequencing #coverage #samtools #depthofsequencing #samflag #sam #bam #alignment #phred #fasta #fastq #singlecell #10X #ensembl #biomart #annotationdbi #annotables #affymetrix #microarray #affy #ncbi #genomics #beginners #tutorial #howto #omics #research #biology #GEO #rnaseq #ngs

[@cibelebandeira2625]

Thank you for this amazing video, slides and explanation!!

[@adepojuayodeji4847]

This was helpful and a good foundation for the mathematical demonstration that is awaited. Hope you get to drop the video of the steps to perform the WGCNA soon. Thanks

[@alirezaebaditabrizi7311]

Excellent, Thank you so much for sharing this tutorial. Hope to see more of you.

[@irodasay3448]

That's what I was looking for, Thank you so much.

[@dyl4nm4rsh4ll]

eloquently explained, thank you!

[@ghadeeralkurdi174]

Fantastic job. Thanks for sharing. It is very helpful.

[@burakkahveci4123]

Thank you for this video, this is perfect video for the introduction

[@aditimehta4886]

Hey Khusbhu! Really helpful video and very well explained. Thank you :)

[@karinchansan]

Very informative video! Thank you!

[@user-db5yx3fi4u]

looking forward to your following video, it is really really nice!

[@Bioinformagician]

It's in the works! :)

[@tulikabhardwaj484]

Thanks a lot . Extremely helpful

[@user-yh6jr5lu4g]

Thank you, this was very helpful!

[@sarrahrose8077]

Thank you for the video - I'm an 18yo learning bioinformatics and this has been a very effective introduction (_:

[@tatianas5637]

Excellent! Thank you!

[@BilalAhmad-gb7ui]

Great effort

[@javierhernando5063]

Thank you so much for the video! Good job! Do you think it could be applied to RTqPCR gene expression values of genes that are known to be "connected"? I would be working with a matrix as well

[@cometellier7983]

Thank you let's start !

[@sijinakp2744]

Thank you Maam.

[@robertoatallah2658]

Thanks very clear

[@pariaalipour61]

Thanks for the helpful video! I was wondering how come batch effect correction is useful when we use the expression matrix as input?

[@ishakyusuf4616]

perfect! thx

[@ajaypradhan5622]

Great video... Just wondering. now it seems using both DEGs and WGCNA is becoming more common. How would you justify this?

[@MGRVE]

Thanks for the video. I keep struggling with the rationale of WGCNA when applied in case-control studies: As typically, in a case-control setting, not that many independent repeats are performed that could produce a meaningful correlation statistic, the overwhelming difference will originate from the original contrast: case-control. Hence, differential gene expression will essentially determine the correlated genes (as the major perturbation is the case-control difference), and thus, differential gene expression is what we are obtaining after all. Dirk

[@Cesar06100090]

Thanks 🎉

[@davidepasini3807]

Nice video! there is some chance to see in future a video on metilation Array analysis?

[@Bioinformagician]

Yes, definitely

[@bergerjuliette617]

Thank you very much for this tutorial!!! could you explain and propose a workflow on weighted topological overlap (wTO) package and or / CoDiNA (Co-expression Differential Network Analysis) ?

[@heresmyname-jz8ft]

any luck with this on youtube maybe?

[@mayconmarcao4554]

loved this theoretical-first approach. It clarified several of my doubts.
Do you think WGCNA also works for DNA methylation? (e.g: filtering features). Thanks.

[@Bioinformagician]

I didn't know WGCNA could be used for DNA methylation until I came across a recently published article. So to answer your question, yes WGCNA can be used for methylation data.

[@yeeshouwwang682]

This was extremely helpful. Would you have any recommended resource on adjusting for batch affects and outlier samples from a data set? The samples I have are not well clustered in PCA and their sample correlations are not well defined either, so it seems adjustments will need to be made. Best.

[@stretch8390]

Have you tried the bioconductor package SVA? You can model known batch effects or unknown through surrogate variable analysis.
What type of dataset also? PCA is only one of many types of dimension reduction

[@Bioinformagician]

Have you tried hierarchical clustering to detect outlier samples?
Another method in addition to PCA to deal with batch effects is ComBat_seq.

[@yeeshouwwang682]

I'll definitely give these methods a a try, thank you for the help!

[@juanete69]

How do you know it's not the opposite direction? How do you know that gene2 is not causing higher gene expression of gene1?

[@yongdongwu8472]

it's very helpful ,and i want know ,can this method can apply to sc-RNA seq data that i got two different treated tissue, or only apply to bulk RNA and mircoarray ,thank you,

[@Bioinformagician]

WGCNA can be applied to scRNA-Seq data

[@whatamess8764]

Thank you for your video. Can we use WGCNA on multiple timepoints? I have 7 timepoints and I'm not sure if it is okay to put them all into one WGCNA analysis or separate out the timepoints and run it separately on each of the timepoints

[@Bioinformagician]

If your goal is to find genes associated with a particular time point then you could perform WGCNA for each time point individually. Another consideration would be how many samples you have, if you have fewer samples for each time point, then constructing one network with all samples make more sense.

[@whatamess8764]

@@Bioinformagician thank you so much! I have a quick follow up. If we have rpkm values, do we log transform and then normalize? Or do we just log transform? If we normalize, do we normalize first or take the log of the values first? Thank you and sorry for bothering you with these questions!

[@md.ishtiakrashid1523]

Is it possible to perform WGCNA analysis from 16S amplicon sequencing data? Although I have done picrust2 prediction. so i am curious.

[@Bioinformagician]

Yes, WGCNA can be performed on 16S amplicon sequencing data.

[@md.ishtiakrashid1523]

@@Bioinformagician Thanks a lot. One more thing is that, shall i use it on raw fastq counts or the taxonomy results (OTU table?). Can you enlighten me please?

[@jaeeponde3508]

hi!! can i download these slides anywhere?