Ask away!
Hello,
I am working with Nanopore cDNA data generated using the SQK-PCB114 kit. I am using Dorado for basecalling and then dorado trim to detect and trim the cDNA primers and extract the UMIs.
I can see that Dorado adds the UMI to the RX tag and the orientation information to the TS tag in my BAM file.
However, I noticed that the UMIs detected by Dorado have different lengths. Here is the distribution I obtained:
23 nt → 2,580
24 nt → 9,534
25 nt → 23,936
26 nt → 39,442
27 nt → 146,937
28 nt → 2,596,358
29 nt → 88,186
30 nt → 15,024
31 nt → 4,522
32 nt → 790
33 nt → 95
Most of the UMIs are 28 nt, but I also observe UMIs ranging from 23 to 33 nt.
For example:
TTTACCATTAATTTCAGCTTACGATTT
TTTGAGCTTCCCCTTGAGCTTCCACTTT
TTTGCCCTTAACCTTACGGTTACGCTTT
TTTCGGATTAAAGTTCCGATTCGGCTTT
TTTAGCCTTAAGGTTGGGGTTCGCGTTT
CTTACCCTTAAGCTTCGAGTTGGCATTT
CTTCCCATTGCGGTTAAGATTAAGATTT
CTTCCGGTTACGCTTAACGTTAACCTTT
CTTCGCATTAAACTTACAGTTACGATTT
CTTGCGATTGCACTTCGAGTTGGCGTTT
CTTGGAGTTAGCATTAAAGTTAGCATTT
GTTAACACTAAGCTTAAGTAGACTAT
My goal is to use these UMIs for PCR duplicate removal. However, this creates a problem for downstream tools such as umi_tools, which expect UMIs to have a consistent length.
I would therefore like to understand whether this variable UMI length is expected behavior for SQK-PCB114/Dorado, or whether it could indicate an issue with UMI detection.
In particular:
What is the expected structure and length of the UMI for SQK-PCB114?
Why does Dorado sometimes report UMIs shorter or longer than 28 nt?
Should the RX tag be used directly as the UMI for downstream PCR duplicate removal?
Is there a recommended way to process these UMIs while preserving the information needed for duplicate removal?
Any advice or information about how Dorado identifies the UMI for PCB114 would be greatly appreciated.
Thank you!
Best regards,
Amal
Ask away!
Hello,
I am working with Nanopore cDNA data generated using the SQK-PCB114 kit. I am using Dorado for basecalling and then dorado trim to detect and trim the cDNA primers and extract the UMIs.
I can see that Dorado adds the UMI to the RX tag and the orientation information to the TS tag in my BAM file.
However, I noticed that the UMIs detected by Dorado have different lengths. Here is the distribution I obtained:
23 nt → 2,580
24 nt → 9,534
25 nt → 23,936
26 nt → 39,442
27 nt → 146,937
28 nt → 2,596,358
29 nt → 88,186
30 nt → 15,024
31 nt → 4,522
32 nt → 790
33 nt → 95
Most of the UMIs are 28 nt, but I also observe UMIs ranging from 23 to 33 nt.
For example:
TTTACCATTAATTTCAGCTTACGATTT
TTTGAGCTTCCCCTTGAGCTTCCACTTT
TTTGCCCTTAACCTTACGGTTACGCTTT
TTTCGGATTAAAGTTCCGATTCGGCTTT
TTTAGCCTTAAGGTTGGGGTTCGCGTTT
CTTACCCTTAAGCTTCGAGTTGGCATTT
CTTCCCATTGCGGTTAAGATTAAGATTT
CTTCCGGTTACGCTTAACGTTAACCTTT
CTTCGCATTAAACTTACAGTTACGATTT
CTTGCGATTGCACTTCGAGTTGGCGTTT
CTTGGAGTTAGCATTAAAGTTAGCATTT
GTTAACACTAAGCTTAAGTAGACTAT
My goal is to use these UMIs for PCR duplicate removal. However, this creates a problem for downstream tools such as umi_tools, which expect UMIs to have a consistent length.
I would therefore like to understand whether this variable UMI length is expected behavior for SQK-PCB114/Dorado, or whether it could indicate an issue with UMI detection.
In particular:
What is the expected structure and length of the UMI for SQK-PCB114?
Why does Dorado sometimes report UMIs shorter or longer than 28 nt?
Should the RX tag be used directly as the UMI for downstream PCR duplicate removal?
Is there a recommended way to process these UMIs while preserving the information needed for duplicate removal?
Any advice or information about how Dorado identifies the UMI for PCB114 would be greatly appreciated.
Thank you!
Best regards,
Amal