How trust4 filters barcodes from cdr3.out and anna.fa to barcode_report?

[yuyuleung]

Dear Dr. Li,

I am looking into the raw outputs from trsut4, like cdr3.out, anno.fa, barcode_report, and barcode_airr, so that I can learn how my data looks like (why there is a little number of complete VDJ). However, I am confused, how the trust-barcoderep.pl and trust-airr.pl actually work to filter invalid contigs and finally got barcode_report and barcode_airr. As I have quick look into these two scripts, I guess that the contigs with parital CDR3 (cdr3_score = 0) , out_of_frame CDR3 (len%3!=0), low cdr3 abundance supported will be filtered. And fianlly the contigs with valid CDR3 should be kept.

However, when I compare the set of barcodes between cdr3.out and barcode_report, I am confused because there are several barcodes that appear to be valid CDR3 contigs but are filtered out from cdr3.out.

To better understand how to process the raw assembly output, could you please explain it more detailed?

Many thanks.
Yuyu

Following file is partial cdr3.out of the barcode 812_669, but this barcode is filtered in barcode_report.
barcode_in_cdr3_not_in_report.txt

[mourisl]

It should not be filtered. I can see the barcode when directly running trust-barcoderep on this file. Is it possible to share the original cdr3.out file? You can send me the file through email (Li.Song@dartmouth.edu).

[yuyuleung]

Dear Dr. Li,

thank you for your answer. I have double check it, it does in the barcode_report. I am sorry about it.

However, can you tell me more about ow the trust-barcoderep.pl and trust-airr.pl actually work to filter invalid contigs and finally got barcode_report and barcode_airr? Cause I have compared the barcode-set after filtering parital CDR3 (cdr3_score = 0) , out_of_frame CDR3 (len%3!=0), low cdr3 abundance from cdr3.out, the resulting barcode-set is not exactly as the set of barcode-report and barcode-airr.

Thank you so much.
Yuyu

[mourisl]

There is an extra filter for gdT. If there are some good abT or B contigs assembled, even without CDR3, those gdT barcode may be regarded as false positives. If you know the barcode that is missing in the barcode report, you can check their chain type in the cdr3.out file.

The trust-airr is based on the barcode_report file, and I don't think it has extra filters.

[yuyuleung]

Oki, I will look into them. Thank you so much :)