I have the RNA sequence data of Euglena gracilis, in 0h, 4h , 24h. Then how can I analyze which genes are osscilatory genes.
Do I have to download the fasta.gz files ? Kindly explain...
•
written
by
ashishkumar30121999 •
1 answer
With just three timepoints you cannot meaningfully do circadian analysis. The crux is to have many timepoints to see whether the data are a good fit for the so-called cosinor model, which is a rhythmic curve. That's pointless with your timepoints. Assuming 24h rhythmicity, imagine your gene have an acrophase (the timepoint of highest expression) at 0h (which is 24h), so it should have a theoretical valley at 12h, but you don't have that timepoint. No circadian analysis with your data. Google cosinor regression for background.
Log in to answer this question.
-
How to calculate fold change values from GSM data
written by teeteen •
-
gene sequence fetching from bam or fastq file of rna seq data
written by fatimarasool135
-
Microarray experimental design and analysis; repeated measures ANOVA, MaSigPro
written by tleona3 •
-
How does differences in library size infuences ChIP-Seq experiments?
written by Assa Yeroslaviz
-
DE analysis with [DESeq2] without replicates
written by john •
-
How to retrieve RNA sequence data with Customprodb
written by genie66 •
-
genes fpkm value of Coexpression module detected by WGCNA
written by l0o0
-
How Does Detection Of Differentially Expressed Genes In Time-Course Microarray Experiments Work?
written by Student •
-
When We Can Carry Out Microarray Meta-Analysis?
written by Cheng Zhongshan
-
How Many Genes Differentially Expressed In Microarray Can Be Seen As Normal?
written by Cheng Zhongshan
As ATpoint noted, finding genes with oscillatory and/or circadian expression patterns with your data set here is not really possible.
What you could do is find the set of those genes that are differentially expressed between
0hand4hAND24hand4hBUT NOT0hand24h. This should give you -- if you are lucky -- some subset of genes that are, in theory, somewhat likely to be varying over time.That won't even do any good. I work on circadian data for years now and can tell you that often we do not find a gene being DE between acrophase and valley timepoint, but it is significantly rhythmic by cosinor. That's just the difference between a goodness-of-fit cosinor model, and the additional rigor that tools like limma and DESeq2 put on. @OP, save the time and don't start circadian analysis of this dataset. Circa is already hard and noise-prone in good datasets, here there is really no point.
I had suggested this assuming that OP is in a position where they can neither generate data nor waste whatever is in their hands. It's also possible that some "results" are being demanded by supervisors who have no inkling of the unsuitability of the data for circadian analysis. In that case, what I suggested would be making the most of a crappy situation. I agree that they might not even find anything that's differentially expressed, but it is at least an analysis that they can do.
I agree with what you've said otherwise, and I generally defer to you in matters of circadian analysis given that I have consulted with you on these issues here on this forum myself.