program killed when creating snap files

[qing5579]

Hello, when I creating snap files by using snaptools, I have met a problem. The bam file was too large (80G) and the program was killed for no reason. When I ran the program on a larger node (RAM 1T), I couldn't produce any valid results, which can be viewed in the H1930001_BS91_IC_002_ATAC_1.snap.qc file.

I found some similar questions before (r3fang/SnapTools#22). It said that spliting the bam file is a good way to solve this problem. I wonder if sanpatac2 still supports direct merging of different snap files.

And, more importantly, is it necessary to split bam files according to barcode? Or is it necessary to make the split bam file contain all the reads of the barcode under a certain number? Can I just split the bam file without considering the barcode information? Thank you.

Here 's H1930001_BS91_IC_002_ATAC_1.snap.qc:

Total number of unique barcodes:             0
TN - Total number of fragments:              0
UM - Total number of uniquely mapped:        0
SE - Total number of single ends:            0
SA - Total number of secondary alignments:   0
PE - Total number of paired ends:            0
PP - Total number of proper paired:          0
PL - Total number of proper frag len:        0
US - Total number of usable fragments:       0
UQ - Total number of unique fragments:       0
CM - Total number of chrM fragments:         0

[kaizhang]

SnapATAC2 doesn't use the snap file format. It uses anndata format. SnapATAC2 can process BAM files with unlimited size as long as they can be stored on your hard drive. You can use this function to generate fragment files from bam files: https://kzhang.org/SnapATAC2/api/_autosummary/snapatac2.pp.make_fragment_file.html#snapatac2.pp.make_fragment_file, and then import them to create anndata objects using import_data.

[qing5579]

I have tried snapatac2.pp.make_fragment_file to process BAM files. But I got BAM files by using snaptools. It seems that the barcode information was missing in the output BAM file.

SRR26212873.1   99      chr10   77185547        42      51M     =       77185892        396     GTCAAATTCAGACCATTAAACCAAAAGGTTGAAGCAGTACTCTAAGGGTTA     FFFFFFFFFFFFFFFFFFFF:FFFFFFFFFFFFFFFFFFFF:FF:FFFFFF     AS:i:0  XN:i:0  XM:i:0  XO:i:0  XG:i:0  NM:i:0  MD:Z:51 YS:i:0  YT:Z:CP
SRR26212873.1   147     chr10   77185892        42      51M     =       77185547        -396    CAGGAACGTGAAGGTGAAGTGAGGTGAGGGAAAGGAGACAGAATACCAAGA     FF:FFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFF     AS:i:0  XN:i:0  XM:i:0  XO:i:0  XG:i:0  NM:i:0  MD:Z:51 YS:i:0  YT:Z:CP

So it is a little bit confusing for me, how do I fill in the barcode_tag and umi_tag parameters? I think barcode_tag="CB" is not suitable, but I don't know what the proper parameter is. Can you give me some advice? Thanks.

snap.pp.make_fragment_file("H1930001_BS91_IC_002_ATAC_1.bam", "H1930001_BS91_IC_002_ATAC_1.tsv.gz", 
                                is_paired=True, 
                                barcode_tag="CB", 
                                shift_left=4, shift_right=-5, min_mapq=30, chunk_size=50000000, 
                                compression='gzip', compression_level=None)

My input fastq file looks like this:

@SRR26212873.1 TTATGGATGCGTGTTAGGCAAT:A00953:497:H5Y7GDSX3:1:1101:1199:1000 length=51
GTCAAATTCAGACCATTAAACCAAAAGGTTGAAGCAGTACTCTAAGGGTTA
+SRR26212873.1 TTATGGATGCGTGTTAGGCAAT:A00953:497:H5Y7GDSX3:1:1101:1199:1000 length=51
FFFFFFFFFFFFFFFFFFFF:FFFFFFFFFFFFFFFFFFFF:FF:FFFFFF

This is the code I use to get the BAM file:

$snaptools align-paired-end	\
	--input-reference=$reference/Homo_sapiens_assembly38	\
	--input-fastq1=H1930001_BS91_IC_002_ATAC_1_1P.fastq.gz	\
	--input-fastq2=H1930001_BS91_IC_002_ATAC_1_2P.fastq.gz	\
	--output-bam=H1930001_BS91_IC_002_ATAC_1.bam	\
	--aligner=bowtie2	\
	--path-to-aligner=$HOME/software/bowtie2-2.5.2/	\
	--read-fastq-command=zcat	\
	--min-cov=0	\
	--num-threads=16 \
	--tmp-folder=./tmp/ \
	--overwrite=TRUE