Hello all,

I'm trying to build a custom CDF (or more precisely, a pdInfo package for use with the oligo R package) for the HTA-2.0 array. Since Brainarray has discontinued creating such files and clear tutorials for building them seem scarce, I'm seeking guidance on using an updated transcript annotation based an updated GTF file.

I have seen a related post on the Bioconductor support site:
Annotation package generation for HTA-2_0 by pdInfoBuilder fails
It discusses the creation of custom CDFs but does not appear to utilize a GTF file for updating transcript annotations.

Files I have downloaded from the ThermoFisher website:

• HTA-2_0.r3.clf
• HTA-2_0.r3.pgf
• HTA-2_0.r3.Psrs.mps
• HTA-2_0_MappingFile.r1.map
• HTA-2_0.na36.hg19.probeset.csv
• HTA-2_0.r3.na36.hg19.a1.transcript.csv

I'm considering using the GTF available from GENCODE Genes.

Which GTF file would be the best to download for this purpose?

• Basic gene annotation
• Comprehensive gene annotation

Could anyone please advise if there is a step-by-step tutorial (preferably from Brainarray https://www.biostars.org/p/49557/) that explains how to:

• Generate a custom CDF using a GTF file for updated annotations?
• Convert the GTF to the appropriate format required by the pdInfoBuilder or similar tools?

I would also appreciate input on the following:

- Should probes be remapped to a newer genome build as part of the process? If so, how this should be done in R or using CLI?

I think if we could make such CFDs it would help others too, for example when you want compare array data with RNA-seq and need to map data using the same annotations.

Any guidance would be much appreciated.
Thanks in advance for your help!

It doesn't appear that Manhong Dai is working at UM any longer, and the webpage appears to have fallen into disrepair (for example, the link to the python scripts is broken now). If you really want a remapped CDF, you will probably have to figure it out for yourself. Or you could try to contact Manhong Dai for help. There's a umich.edu email you could try.