search

JiekaiLab / scTE

MIT License
87 stars 27 forks source link

Work with Seurat and bam files merge #16

Open Counts-Xin opened 3 years ago

Counts-Xin commented 3 years ago

Very nice work! Now I have two bam files from the cellranger (possorted_genome_bam.bam), how can I merge those two bam files with scTE, AND because I have run the Seurat before scTE, how to transfer the analysis results from Seurat like cell ananation and cell emmbeding to scTE.THANKs

jphe commented 3 years ago
  1. You can provide with multi bam files, scTE treat them as one sample and merge them; Or you can run scTE seperately and then merge the results by seurat or scanpy.
  2. scTE output csv file by default, you can load it into seurat and can tranfer with your old seurat results.
Counts-Xin commented 3 years ago

yeas, thanks you, but when I run scTE,the error happened:

ERROR : The input file 200108J_E10_AGM/outs/possorted_genome_bam.bam has no cell barcodes information, plese make sure the aligner have add the cell barcode key, or set CB to False。

and my bam file looks like this:

samtools view 200108J_E10_AGM/outs/possorted_genome_bam.bam

A00268:623:HYJNGDSXY:3:2178:10863:28823 1040 1 4491510 255 150M * 0 0 GTGATTGTGGGGAGCAAGTCCCTCAAGGCATTTAAAACAAAAATCTCGTGTAGCCCCTCAACTGTTCAAGTGGCAGACAAAATAAATTACCATAAATTATATGCCAAGGCAACTTAAAAACAAAACAGCTTGACGCTAAGGATTTCCTTA FFFFFFFFFFFFFFFFFFFFFFFFFFFFF:FFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFF,FFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFF:FFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFF NH:i:1 HI:i:1 AS:i:148 nM:i:0 RG:Z:200108J_E10_AGM:0:1:HYJNGDSXY:1 TX:Z:ENSMUST00000027035,+2398,150M;ENSMUST00000192650,+2972,150M;ENSMUST00000195555,+1567,150M GX:Z:ENSMUSG00000025902GN:Z:Sox17 fx:Z:ENSMUSG00000025902 RE:A:E xf:i:17 CR:Z:AGGCATTCAACCGATT CY:Z:FFFFFFFFFFFFFF:F CB:Z:AGGCATTCAACCGATT-1 UR:Z:TGGGCCTATTCC UY:Z:FFFFFFFFFFFF UB:Z:TGGGCCTATTCC

Does I need to change some parameters?

jphe commented 3 years ago

Can you check if your bam file has any reads without barcode and UMI tags? like this https://github.com/JiekaiLab/scTE/issues/5

Counts-Xin commented 3 years ago

Thank you, this problem has been resolved, but it runs very slow, and I have set the workers as 20, Can I make it faster?

jphe commented 3 years ago

I think it may run out of memory, you can set with fewer thread, scTE takes ~10Gb memory each thread for human and mouse genome.