search

ldiao / MixMir

A mixed linear model approach to small RNA motif discovery
4 stars 4 forks source link

Understanding output #2

Open 18209018 opened 8 years ago

18209018 commented 8 years ago

Hello, I got the result and I have a question. So, from this program, I can get the miRNA which is predicted to bind the mRNA based on the log fold change of genes. The output file is a CSV file of miRNA matched to a motif with p-value ranked from the lowest to a certain number of rank. Is it possible to know what gene are those miRNA bind to? All I can get is the rank, motif, p-value, coef, NUTRs, and miRNA matched. I want to know what gene paired with what miRNA. I attached my result file. Thank you. BDC_647-MixMir-results.txt

ldiao commented 8 years ago

Hi Bharata,

MixMir does not return a list of genes that the miRNA binds to: it searches over all possible motifs and finds ones which are highly associated with changes in gene expression values, and thus are more likely to belong to active miRNAs. Thus, any gene whose UTR contains the motif is more likely to be targeted.

Hope that helps, Liyang

On Thu, Mar 3, 2016 at 12:58 AM, Bharata kalbuaji notifications@github.com wrote:

Hello, I got the result and I have a question. So, from this program, I can get the miRNA which is predicted to bind the mRNA based on the log fold change of genes. The output file is a CSV file of miRNA matched to a motif with p-value ranked from the lowest to a certain number of rank. Is it possible to know what gene are those miRNA bind to? All I can get is the rank, motif, p-value, coef, NUTRs, and miRNA matched. I want to know what gene paired with what miRNA. I attached my result file. Thank you. BDC_647-MixMir-results.txt https://github.com/ldiao/MixMir/files/156229/BDC_647-MixMir-results.txt

— Reply to this email directly or view it on GitHub https://github.com/ldiao/MixMir/issues/2.

18209018 commented 8 years ago

Thank you Liyang for your reply.

Another question about the search of the motifs. The input is FASTA for the 3'UTR for the gene and the miRNA. I imagine it searches by "aligning" the miRNA sequences to the 3'UTR sequences right? Probably I can modify the code to inform me the gene which miRNA motifs sequences is found. I imagine the result would be a pair of miRNA and gene name. How do you think?

ldiao commented 8 years ago

Hi Bharata,

Yes, you can certainly do a search for all genes which contain the motifs of interest (the column NUTRs tells you how many such genes exist in your data). In that case, you can obtain a list of miRNA-gene pairs as desired, and maybe sort by those genes with greatest change in expression.

Best, Liyang

On Thu, Mar 3, 2016 at 2:32 PM, Bharata kalbuaji notifications@github.com wrote:

Thank you Liyang for your reply.

Another question about the search of the motifs. The input is FASTA for the 3'UTR for the gene and the miRNA. I imagine it searches by "aligning" the miRNA sequences to the 3'UTR sequences right? Probably I can modify the code to inform me the gene which miRNA motifs sequences is found. I imagine the result would be a pair of miRNA and gene name. How do you think?

— Reply to this email directly or view it on GitHub https://github.com/ldiao/MixMir/issues/2#issuecomment-191927138.