Using different adapters across samples

[bschilder]

Hello @oscarwilkins @CharlotteAnne @jernej-ule,

Thank you sooo much for this amazing tool! Really appreciate how well designed it is.

We're trying to de-multiplex some single-cell histone mark sequencing data our lab has generated using a new protocol. However in our case, we have slightly different adapter sequences for each sample (cell), in addition to the unique barcodes. @Al-Murphy

From my understanding, Ultraplex only takes 1-2 adapter sequence for all samples (forward and reverse). I'm wondering if it's possible to specify different adapter sequences specific for each sample.

Many thanks in advance,
Brian

[Delayed-Gitification]

Hi Brian,

Thank you for your message. Would this be for single or paired end data?

[bschilder]

Paired end data, sequenced on NovaSeq 6000

[bschilder]

So far i've tried writing a loop that iterates over each adapter sequence and tries demultiplexing each cell one at a time. but obviously this is a very inefficient approach that requires scanning through the entire fastq file many times. Plus I'm not sure if it would even yield the same results as running ultraplex the way it's supposed to be used.

I'm wondering, would somehow concatenating the barcode with the adapter sequence and supplying both a the "barcode" work? this makes some assumptions that may be wrong though. @Delayed-Gitification

[oscarwilkins]

Hi Brian, I think the easiest thing may be to run ultraplex ignoring adaptors, using a setting that returns the original untrimmed read (experimental), then run cutadapt on each demultiplexed fastq. I can send you an experimental version that writes out fastqs without the adaptors trimmed. How does that sound?

…

On Mon, 30 May 2022, 18:34 Brian M. Schilder, ***@***.***> wrote: So far i've tried writing a loop that iterates over each adapter sequence and tries demultiplexing each cell one at a time. but obviously this is a very inefficient approach that requires scanning through the entire fastq file many times. I'm wondering, would somehow concatenating the barcode with the adapter sequence and supplying both a the "barcode" work? this makes some assumptions that may be wrong though. — Reply to this email directly, view it on GitHub <#36 (comment)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AOCA3HLJIJ37AX4TRSL4E6DVMT323ANCNFSM5W5HQN6Q> . You are receiving this because you were mentioned.Message ID: ***@***.***>

[Delayed-Gitification]

This should be very fast - cutadapt on the demuxed fastqs can be run in parallel on a cluster so will require very extra time

[bschilder]

@oscarwilkins that makes quite a bit of sense. And yes, I'd definitely like to try out the experimental version. Thank you so much! Here's all my contact info, in case you need it: https://bschilder.github.io/BMSchilder/#contact

[Delayed-Gitification]

Hi @bschilder I've published a new branch called "experimental_no_trim", which has a new argument "--skip_adapter_trim".

I'm not easily able to test right now as I am away, but you are welcome to give it a try. Let me know if there's any issue!

[bschilder]

ok great, thank you! i'll give it a try and let you know if I run into any issues (tho don't feel the need to reply until after you're back)

Thanks again

[Delayed-Gitification]

To remove the inline barcodes in the mate read you might need to use some N bases in cutadapt.

Eg if the in-line barcode in read 1 is 5 bases long, and the adapter for read 2 is "ATGCATTG", then for trimming read 2 you should use "NNNNNATGCATTG" when running cutadapt