Entering edit mode
3.5 years ago
nkabo
▴
80
Dear all,
I have a list of genes, composed of over 19 000 rows. I would like to give this list as an input and get which tissues or cell types that these genes are enriched at most and at least as an output. Ideally, I would like to see which genes are enriched in which cell-type or tissue-type. Is there a R package for such analysis? Or could you suggest me another tool? Thanks in advance!
Can you be a bit more specific about the type of data you have? I.e. 19,000 genes sounds like it's a fairly comprehensive list, e.g. of all protein-coding genes in mouse or humans. No cell type that I know of expresses all of these genes. Ergo, I assume you have some sort of expression metrics associated with that list, but I do not know. Please share what type of experiment was used to generate the data and what format your expression values (or whatever types of values you have).
If this is from bulk RNA-seq, xCell might be of use.
Thank you for your response. I have had fastq files from RNA-Seq, the final result file is a txt file involved gene names, samples and expression values. The data is from endothelial cells, so I would like to exclude genes that:
in order to eliminate the tissue-specific properties.
So you want to get rid of genes that are differentially expressed in relation to "other types of cells"? How many other types of cells? How are you going to compare to each type of other cell? What you're asking is extremely broad, and you likely need to better define exactly what you want out of this analysis.
Alternatively, you may be able to find some info you need from resources that already exist, like the EGA Expression Atlas, the Human Cell Atlas, or the Human Protein Atlas.
Thank you for your reply. I will check Human Cell Atlas. The reason why I would like to find cell-type specific genes is that I have RNA-Seq data from 2 different cell types (one is from leukocytes and other is from endothelial cells) and I would like to compare these 2 datasets. Since there is lack of sufficient number of controls, I try to eliminate the effect of coming from different cell types. Therefore, I aim extracting such genes which are for example, highly specific to leukocytes and highly specific to endothelial cells. Due to lack of samples, I had to look from a different approach, but I do not think it is much plausible. Could you suggest me a better way to eliminate cell-type effect? Thanks
What exactly are you trying to gain from that comparison if not tissue-specific differences?
Data for case and control samples are from different types of cells (case is from leukocytes and control is from endothelial cells). I have to compare case and control for differential gene expression analysis. Due to low number of controls and cases, I have only these two samples (1 for case and 1 for control), and I was asked to compare them to obtain differentially expressed genes. Therefore, I am trying to extract the cell-type specific effect on differentially expressed genes as possible. I aim to sort of "normalize" the data by removing genes which are cell-type specific. Therefore, after comparing gene expressions of case and control, I would like to remove genes which has leukocyte-specific and endothelial cell-specific expressions.
While I understand the logic behind that, you should realize that such an analysis will likely be plagued by false negatives/positives and any results will be very hard to trust. Additionally, it would likely garner strong negative reactions/comments during review. If at all possible, I would attempt to acquire more suitable controls.
Thank you for your reply, I will try obtain control data coming from leukocytes.