Easy question from a noobie

[HESmith]

It looks like a 2D graph, so I'm reasonably certain it's comparing only two of the replicates. Each point represents the read counts for a single gene in the two replicates.

[kwatts59]

I think I figured it out.
According to the paper, they performed 5 technical replicates on a brain sample.
That means they performed the RNA-seq library build and sequencing on the same brain sample five times.

So for example, if gene X has the following number of reads that aligned
Replicate 1, 152 reads aligned
Replicate 2, 127 reads aligned
Replicate 3, 98 reads aligned
Replicate 4, 173 reads aligned
Replicate 5, 111 reads aligned

The following 20 points are plotted on the graph
(152,127),(152,98),(152,173),(152,111)
(127,152),(127,98),(127,173),(127,111)
(173,152),(173,127),(173,98),(173,111)
(111,152),(111,127),(111,98),(111,173)
(98,152),(98,127),(98,173),(98,111)

So basically this graph shows that the fewer reads that map to a gene, the less reliable the data.

[HESmith]

Is your question "what is a technical replicate"?

[kwatts59]

Can someone please explain figure 2 "Scatterplots of Technical Replicates of Brain Sample" in the paper below?
The x and y axis are both labeled the same "Number of aligned reads (log scale), Brain".

http://www.illumina.com/Documents/products/whitepapers/whitepaper_RNASeq_to_arrays_comparison.pdf