GATK version used ?

[divyanshusrivastava]

Do you have a log of the exact GATK version that was used when building the align script ?

[AAnnan]

gatk-4.6.0.0

[divyanshusrivastava]

I use 4.5.0.0

Sometime a get a warning like this https://www.biostars.org/p/353352/
I solved it earlier using

# Mark duplicates
echo "Marking duplicates..."
${PathGATK} MarkDuplicates -I ${RGID}_TEMP.bam -O ${RGID}_TEMP_NoDup.bam -M ${PathOutputPicardDupStats} --REMOVE_DUPLICATES ${RemDups} --BARCODE_TAG RG --MAX_RECORDS_IN_RAM 10000000  --READ_NAME_REGEX null --ASSUME_SORT_ORDER queryname

Do you think its correct ?

[AAnnan]

If you don't care about subtypes of read duplicates (optical vs PCR), you shouldn't care about this.

…

________________________________ From: Divyanshu Srivastava ***@***.***> Sent: Friday, March 21, 2025 11:55:43 AM To: AAnnan/scPreproc ***@***.***> Cc: Ahrmad Annan ***@***.***>; Comment ***@***.***> Subject: Re: [AAnnan/scPreproc] GATK version used ? (Issue #4) I use 4.5.0.0 Sometime a get a warning like this https://www.biostars.org/p/353352/ I solved it earlier using # Mark duplicates echo "Marking duplicates..." ${PathGATK} MarkDuplicates -I ${RGID}_TEMP.bam -O ${RGID}_TEMP_NoDup.bam -M ${PathOutputPicardDupStats} --REMOVE_DUPLICATES ${RemDups} --BARCODE_TAG RG --MAX_RECORDS_IN_RAM 10000000 --READ_NAME_REGEX null --ASSUME_SORT_ORDER queryname Do you think its correct ? — Reply to this email directly, view it on GitHub<#4 (comment)>, or unsubscribe<https://github.com/notifications/unsubscribe-auth/AO5EE27KON2RMBY4ETSZTR32VPV27AVCNFSM6AAAAABZPVRSGCVHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMZDONBTGAYTINBSGQ>. You are receiving this because you commented.Message ID: ***@***.***> [divyanshusrivastava]divyanshusrivastava left a comment (AAnnan/scPreproc#4)<#4 (comment)> I use 4.5.0.0 Sometime a get a warning like this https://www.biostars.org/p/353352/ I solved it earlier using # Mark duplicates echo "Marking duplicates..." ${PathGATK} MarkDuplicates -I ${RGID}_TEMP.bam -O ${RGID}_TEMP_NoDup.bam -M ${PathOutputPicardDupStats} --REMOVE_DUPLICATES ${RemDups} --BARCODE_TAG RG --MAX_RECORDS_IN_RAM 10000000 --READ_NAME_REGEX null --ASSUME_SORT_ORDER queryname Do you think its correct ? — Reply to this email directly, view it on GitHub<#4 (comment)>, or unsubscribe<https://github.com/notifications/unsubscribe-auth/AO5EE27KON2RMBY4ETSZTR32VPV27AVCNFSM6AAAAABZPVRSGCVHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMZDONBTGAYTINBSGQ>. You are receiving this because you commented.Message ID: ***@***.***>

[divyanshusrivastava]

Ok. But why do you think the warning arose in the first place?
Using GATK4.5 I get this warning even on the samples used earlier.

[AAnnan]

Could you share the error message? or the fastq record in question? I suspect either we used a different sequencer that GATK might not be used to yet (ex. AVITI), or the way I add the barcodes to the read name is freaking out the regex used for optical duplicate detection.

At any rate, this should not affect duplication detection. Only tagging of the type of duplication. And potentially library size estimation. But I don't think we were using either of these infos before.

If you'd like to fix this properly, MarkDuplicates has documentation on how to use the --READ_NAME_REGEX parameter: https://gatk.broadinstitute.org/hc/en-us/articles/21905036102043-MarkDuplicates-Picard#--READ_NAME_REGEX