merging per-transcript data for form per-gene data

Question

[missnc]

Hi!

I have a set of 43,000 genbank accession numbers (NM_000014, NM_000015, NM_000016, NM_000017) for human transcripts. I am trying to run GENIE3, for which I believe I need the data to be in a per-gene rather than a per-transcript format. I am looking for a way to merge the microarray data I have for each transcript to produce a gene-by-sample data frame with expression data for each gene. I don't mind much what method is used to find the expression level from the expressions of each transcript - I am primarily looking for a way to identify transcripts that refer to the same gene and merge them.

Thank you!