Barcode assign without using CAMPAREE

[dcansugurer]

Hello,

I am trying to use BEERS2 without running CAMPAREE beforehand. I manually generated molecule_file.txt for my samples and provide the i5 and i7 barcode sequences in the config file. I can generate the reads at the end but only the unidentified FASTQ files have reads, the others are simply empty.

I double-checked my sample numbers/names and barcode assignments seems correct in the config file, yet I couldn't somehow assign them to the output FASTQs.

Should I run CAMPAREE to achieve this or is there any other thing that I do manually?

EDIT: I checked the i5+17 barcode pairs in the headers of the output reads and every single read has its own, random I assume, i5+i7 barcode pair. Probably that is why demultiplexing fails. It seems somehow the output read headers do not have the barcodes I give in the config.yaml file.

Thank you very much in advance.
Best regards.
Cansu.

[tgbrooks]

Hi Cansu,

Can you please provide your configuration file? Ideally also a small set of your molecule file (maybe first 10 rows). If you don't want to share them publicly here, can you please email them to me at thobr@sas.upenn.edu? Thanks!

Off the top of my head, there's a chance that your config is set unusually in such a way that sequencing errors are very high, which includes the barcodes, and so the barcodes end up appearing random.

[dcansugurer]

Thank you very much for your reply. I am attaching the config file and first 10 rows.

Actually sequencing errors probably are very high because for the beginning I have a very limited set of regions that I want to simulate (I have 100 unique molecules). And my molecules cannot be retained until the end of the run with the default settings, eventually the pkt files become empty and I am having counter error. To avoid this, I modified the config file to be really permissive. Then ended up having reads but they are randomly barcoded.

I can increase the number of molecules in my molecule file but I still want every region to be included without losing them due to sequencing errors. ( I am assuming it is because of the sequence somehow) I don't have much experience in read simulation so I am little confused :)

Thanks again.
Best regards.
Cansu.

config.yaml

molecule_file_first10.txt

[tgbrooks]

Hi, it looks like you've essentially disabled every step, including the bridge amplification. What this means is we end up with just 2 molecules in each cluster which isn't enough florescence signal to overcome the background. To solve this, increase cycles parameter of bridge_amplification_step.BridgeAmplificationStep back up to 10.

It sounds like the reason you did this was that you didn't get enough output originally. In BEERS, the input is expected to be a realistic sample of RNA molecules, which means it needs huge numbers of input molecules. It sounds like you're giving BEERS just 100 molecules per sample. Input molecules don't have to be distinct from each other (just as a real RNA sample will have many copies of the same transcript), so you can just duplicate your 100 molecules 1000 times each and see what you get.

Best,
Tom