Different splice abundances than kb tutorial

[carbycrab]

Hello,

Thank you for creating such an awesome package for RNA velocity matrix. I am trying to replicate the tutorial posted here, but I am getting different results than described in the tutorial. I am successful downloading the data, downloading the pre-built index, and generating the RNA velocity matrix. However, when I concatenate adata_06 and adata_07, the dimensions of my data are:

AnnData object with n_obs × n_vars = 807273 × 58367
    obs: 'run', 'bcs', 'batch'
    layers: 'spliced', 'unspliced'

instead of

AnnData object with n_obs × n_vars = 374814 × 58367 
    obs: 'batch', 'bcs', 'run'
    layers: 'spliced', 'unspliced'

Additionally, when I run the splice proportions for adata, I get: Abundance of ['spliced', 'unspliced']: [0.5 0.5]. I tried generating the RNA velocity matrix for some of our own 10x datasets, and also get 50/50 splicing proportions. Here is the code I ran:

kb count --h5ad -i index.idx -g t2g.txt -x 10xv2 -t 5 -o SRR6470907 -c1 spliced_t2c.txt -c2 unspliced_t2c.txt --lamanno SRR6470907_S1_L001_R1_001.fastq.gz SRR6470907_S1_L001_R2_001.fastq.gz SRR6470907_S1_L002_R1_001.fastq.gz SRR6470907_S1_L002_R2_001.fastq.gz 

I removed the flag --filter bustools since it produces this error: kb count: error: argument --filter: not allowed with argument --lamanno.

Does anyone know why I might be getting those splice percentages and what I can do to fix this issue?

Thanks so much!