assignment part 2 mark

[farnushfarhadi]

Question	Grade	Comment
6.1	1.75/2	For PCA, it'd be the "x" matrix that you're interested in, not "rotation".
6.2	1.75/2	colnames are not correct.
6.3	0.5/1	Missing a graph. saved your table in your local computer
7.1	1.0/1
7.2	0.75/1	edger.deep.results does not have columns as stated in question. saved in your local computer
7.3	0.75/1	did not packaged the results
7.4	1.0/1
8.1	1.25/2	It is not clear how you choose genes for different parts of the questions. YAR003W is not DE only in rna sequencing data! Also, you could have two platforms in one plot to make the comparisons easier.
Mechanics	4.75/5	things to improve: comenting different parts of your code. having your document to be more organized. Working on graphs like having titles for them. Could use more narrative on the interpretation of figures
Late?	0.0
(Total)	13.5/16

[fangwuwang]

@farnushfarhadi Thanks for the grading. I just want to mention that for the three lists of genes in question 8.1, there was an explanation with the venndiagram just before the 8.1 part, which stated that

q1: common genes of microarray and sequencing data

q2: only identified from sequencing data

q3: only identified from microarray data

so the genes were picked from the q1, q2, q3 lists respectively using the sample() function. Does that sound reasonable? Thanks! Fangwu

[farnushfarhadi]

Hi @fangwuwang

It is hard to tell which gene is DE according to your plots. One other thing is that when you want to compare a gene in array and RNASeq data, you want your plots have same scale.

when you use sample to choose a gene with non significant q value (> 0.05) , it can choose a gene with the lowest q value which is greater than 0.05. So the q value is still very low but it is only marginally greater than 0.05 so you do not call it signifiant any more.

The safest way to find a gene which is DE in RNASeq data but not in array, is

• filter genes with q value > 0.3 [something big! ]
• filter genes with q value < 0.05
• get the intersect

[fangwuwang]

Thanks for the comment @farnushfarhadi
Just wanted to clarify how I came up with DE genes in the Venn graph and the q1,q2,q3 lists. I used a tool for presenting the venn graph called venndiagram which calculates the overlap between two list of characters using "x=" and "y=" arguments. In this case "x=deg_deep" which is the list of genes from the RNA-seq DE table with FDR<1e-5 (see 7.4), and "y=array" which is from microarray DE table with FDR<1e-5 (see 6.3). Sorry for the poor annotation of my codes.