Replication in PCR-array experiments

Question

[aaozmen]

Is it enough to use only three biological replicates for correct statistics PCR-array analysis or three biological replicates and two technical replicates or one biological replicate and three tech replicates?

Our study divided into a total of 3 main groups and 15 subgroups. Each main groups contains 5 subgroups. These subgroups are associated with time (1h, 2h,3h, 4h and 5h). Like this:

Group A: Control                                                     Group C: Treatment Group
subgroup I, (1h)...........3 rats                                    subgroup I, (1h)...........3 rats                    
subgroup 2  (2h)...........3 rats                                    subgroup 2  (2h)...........3 rats
.....                                                                .....
subgroup 5  (5h)...........3 rats                                    subgroup 5  (5h)...........3 rats

Group B: Negative Control
subgroup I, (1h)...........3 rats
subgroup 2  (2h)...........3 rats
.......
subgroup 5  (5h)...........3 rats

My questions

3 rats in each group are biological replicates. (I think). Are biological replicates enough? Or are technical replicates necessary? How many is enough for technical replicates?

and

At the end of the project, I will have results of each animals' 88 gen expression. Located timing in each sub-group is important, too. What program should I use to compare differences in gene expression? And how can create a heat map?

Answer 1

[Carlo Yague]

Biological replicates can be enough if the only question you are interested in is identifying DEG, i.e, comparing the variability between conditions with the variability between biological replicates (that includes both technical and biological variability).

But technical replicates allows to differentiate technical from biological variability and to know if something went wrong during the experiment. So technical replicates are not as essential as biological ones, but since it costs you less (in term of work, money and stuff), you might as well includes them. They can also increase the power of your tests to identify DEG. How many technical reps is up to you, you need to find the balance between cost and benefit. I wouldn't have more than three though.

To find DEG, you can use "R" or perhaps Excel.

For heatmap, you can use "R" or cluster with cluster and treeview.