I am using the Developmental Transcriptome tool from the Allen Brain Atlas to determine at what stages of development a specific gene is expressed in the human brain.
I have read the official documentation as to how the RNA-Seq data for the Developmental Transcriptome was generated and analysed. In the documentation, it says that they used postmortem human brain specimens from a range of different developmental stages (embryonic, early prenatal, early mid-prenatal, late mid-prenatal, late prenatal, early infancy, late infancy, childhood, adulthood).
The heatmap showing the gene expression levels across the different brain samples was generated from the RNA-Seq data. I know that heatmaps show the gene expression levels in a number of samples (e.g. treatments, disease) relative to a control sample.
However, when showing the gene expression levels across the different brain samples they used, I was wondering if there is a reference brain sample that was used? E.g. they are comparing gene expression in other brain samples relative to the reference brain sample (e.g. brain sample from youngest donor). I have read the documentation and there is nothing mentioned regarding this.
I was wondering if anyone who has used the Human Developmental Transcriptome tool before knows whether the gene expression levels are compared against a reference brain sample? Any advice is appreciated.
Hi @ceno980. These values are RPKM (reads per kilobase per million) normalized and then converted to a log2 scale. This is done separately for each sample and therefore there is not a reference sample. When performing a differential search (as below), the contrast structure(s)/stage(s) can be considered the reference to which the fold change is calculated. Please reply to this post if you still have questions.
Thank you for your reply. I have downloaded the RNA-seq data for a gene of interest from this page.
I would just like to confirm that the gene expression values when you download data for a particular gene from the Developmental Transcriptome page showing the heatmap are in the units log2 RPKM (since this is what is said in the text file explaining the data).
I would like to clarify this as if you download the Developmental Transcriptome data for all genes (RNA-Seq Gencode v10 summarized to genes) from this page, in the text file accompanying the Excel spreadsheets, it is said that the gene expression values are in the units RPKM.
For any heatmap displayed on our atlases (including the one you’ve shown) the downloaded value corresponds to the value shown in the heatmap, which in this case is log2 RPKM (in other atlases it can also be z score). As for the download page, the values are whatever the description says (in this case RPKM). You can quickly confirm whether a table downloaded from a “download” tab is log2 normalized by checking the maximum value–if it’s 10-20, then it’s log2 normalized, but if it’s 100-10,000+ then it’s linear scale (this is not necessarily true for a heatmap with only a few genes shown).
