PacBio data analysis help requested...

[Lmathew]

I am a newbie into PacBio sequencing. I currently have data from a targeted sequencing study of a specific transcript.

What we have is data from 3 groups of individuals (10-homozygous resistant, 10-susceptible and 10-heterozygous for gene A).cDNA of transcript for gene A was PCR amplified for each individual and barcoded specifically for each individual (total 30 unique barcodes). We then did PacBio sequencing on the PCR product after pooling it. Raw data produced reads that were expected size and also huge number of shorter reads half the size.

I am now trying to figure out what the best approach is to analyze the data in order to obtain all variants forms of this transcript A from each individual.

Would it be better to run LAA or ROI analysis after splitting the reads based on barcodes? Also, how do I capture all the reads for this analysis (as shorter reads seems to be fragments of this transcript).

Please help...thank you!

[rhall]

How long is the PCR product? LAA would be the recommendation if the insert size is too long to get a high CCS (ROI) yield >3kb ish.
Do the shorter fragments have both barcodes? How were the barcodes added, barcoded adapter, universal primers (two step PCR) or simply added to the PCR primer sequence?

[Lmathew]

Thank you for your response!

PCR product from each sample was ~ 5 kb and unique PacBio barcode was added directly to gene specific primer and each sample cDNA was PCR amplified using specific barcoded gene specific primer. The PCR fragment (~5 kb) was gel purified and pooled after normalization. PacBio reads obtained had substantial 2 kb reads and expected ~ 5 kb reads.

On further analysis of the shorter reads, they seem to be varying fragments of the PCR product. Some reads with barcodes and some internal fragments of the PCR product (so not barcoded and not sure how these came up).

What should I do to capture all variants forms of this transcript A from each individual?
Thanks again!

[rhall]

LAA should be able to deal with the fragmented data, so long as there are reasonable number of full length sequences.

[Lmathew]

Thank you, rhall. I do have decent full length reads. I appreciate your help.