In need of help with grad school genomic analysis project

Hello everyone,

I am looking for conceptual help with a grad school genomic analysis project. Instructions:

Select one or more papers, articulate a new hypothesis, and use the dataset(s) to support (or refute) that hypothesis.

Perform at least two analyses/visualizations on two different data models (RNA-seq, microRNA, copy number, ChIP-seq, etc), or apply at least two different methods to test your hypothesis.

Integrate the different analyses to enhance your conclusion(such as combining microRNA and mRNA data or using pathway analysis/visualization algorithms to confirm functional similarity).

I had planned to take RNA-seq data from a mouse study, and extrapolate the impact of intermittent fasting-induced differential gene expression to cancer survival in humans. I knew would be stretch and more of a fun, speculative approach. The plan was to get the mouse RNA-seq data, perform differential gene expression, convert the mouse gene ID to human gene ID, as to pretend that this was a human study from the beginning. From here I would perform gene set enrichment analysis to find enriched biological pathways. If I were to find any, I would then perform survival analysis with cancer patient mRNA expression data, to find any associations with intermittent fasting up/downregulated genes form the identified enriched gene sets.

I presented my project idea to a professor and he was not the biggest fan. Although the biological implication might be a far stretch, I assumed that it would be an acceptable since I THINK it follows the instructions. I am hoping to get some suggestions for how I could either improve my approach, or a general framework of analysis that would follow the instruction provided, to then look for a paper(s) with data sets that I could use. For my personal interests, I would like to focus on the metabolic side of human longevity/age-related diseases, I just lack the knowledge of how to do so, with these instructions.

link of paper I initially wanted to use. - https://www.ncbi.nlm.nih.gov/pmc/articles/PMC6764061/pdf/10.1177_1559325819876780.pdf

I appreciate any help, thanks in advance!

project gradschool anaylsis data

• 2,288 views • link

updated 5.3 years ago by qmarulfiz

• written 5.3 years ago by lopez52891 •

2 answers

Did he say why he wasn't a fan? Did he give you any guidance or recommendation? It sounds to me like you're reaching un-necessarily too far for this project. I don't see why you need to cross organisms to pursue your stated interest, and I think you could probably drum up some more, focused ideas by spending more time perusing the literature. Genomic data sets are great for hypothesis generation. Advanced searches in GEO might also help you find ideas to pursue your interest.

• 0 views • link

written 5.3 years ago by seidel

From what I understood is that this project is just to show our knowledge of data analysis in R. He wasn't very clear, we both were in a rush today, he just we implied that he did not like it because the connection of GSEA results to patient cancer data would be weak. Which i agree its a leap, but perhaps not too farfetched, unless there is something very obvious that I am missing.

Oh i agree that it might be overkill, but everything that I have explained mirrors was we did in class or what I have worked on for my master project, so most of the coding is done. The conversion to human gene ID is to show technical ability in R, and to pretend that the originally study was done in humans, so I could interrogate patient gene expression as well.

Yes I agree, the GSEA would be the exploratory data used to formulate my hypothesis. As much as I would like to go through the literature, this assignment doesnt really warrant that amount of effort! I assumed what i purposed would suffice in showing my skills while still having an interesting presentation. I am just curious if there is something blatantly obvious wrong in my thought process, that I am missing.

Thank you for the response!

• 0 views • link

written 5.3 years ago by lopez52891 •

I am new in bioinformatic (just involve for about 3 month)

Everything you wrote till the GSEA analysis is similar to my initial plan for the 1st half of my thesis. From there I can compare the significant genes with MS/MS protein ligand database to see what ligand can affect that gene protein.

The motivation for doing it also similar: 'to showcase R data analysis to my professor'. My professor is an old school biochemist that relies alot on western blot and MS/MS metabolites measurement.

I try to present the idea to my assistant professor (which also old, but since he doing genetic, he kinda understand abit bout these stuff, but he still not convince i can persuade my professor to allow me to do this).

My objective is to see the expression pattern of list of genes (which was actively studied in our lab) on various diseases or genetic disease model dataset. Meta analyses is the method I read that is required for this analysis. But how to actually do it is not within my grasp.

It will be good to hear more from you and how you progress with this plan.

• 0 views • link

written 5.3 years ago by qmarulfiz

Thank you for the reply. Yeah I am "relatively" new to this but do have a couple years of experience. When you say thesis do you mean for a master's program or are you pursuing a phd?

Honestly if you have free time maybe we can set up a zoom call where we can chat about your idea, I dont now how much help I would be, but I can provide my opinion/suggestions which can hopefully help, especially since you are new to bioinformatics.

In my experience, "something blatantly obvious wrong in my thought process, that I am missing", was the main reason why some data analysis propositions were turned down from my mentor. That being said, its the deeper knowledge of tying data analysis to what biological implications we can get from the computation. I guess we can chat about it later on if we ever meet up virtually, but why exactly did the "middle man" think that your mentor would deny you permission.

• 0 views • link

written 5.3 years ago by lopez52891 •

Sure!

I would love to have somebody to discuss or talk about my idea properly. I have been doing this blindly without guidance and I don't actually have time to experiment or do this on trial and error basis (PhD thesis due in 6 month!).

My knowledge on this topic is still kinda shallow though. Not sure if I can understand most of the stuff even if you give suggestion. I drop you my email once I figure out how to send my email to you privately.

• 0 views • link

written 5.3 years ago by qmarulfiz

Here is my email, I think I write here will be much faster.

qamarulfiz@gmail.com

Drop me an email and then we can arrange zoom call later.

• 0 views • link

written 5.3 years ago by qmarulfiz