Open alexpan82 opened 4 years ago
Hi Alex, As long as your cells are B-lineage cells, this low mapping rate is very strange. It would help me to know a bit more about your dataset and setup. Which version of Bowtie are you using? Do you provide all reads to BraCeR, or only reads that mapped in your HISAT2 alignment? Did you run BraCeR separately for fastq reads for each cell? Which parameters did you run bracer assemble with? It could also help if you provide example input and output for one failing and one passing cell.
Best regards, Ida
Ida,
Thanks for your response and sorry for the delay. 1) Bowtie 2.3.0 2) I provide pre-trimmed reads (AdapterRemoval) to bracer. There was no HISAT2 filtering 3) I run bracer separately for each single cell/fastq file 4) bracer assemble --ncores 12 --config_file $PATHTOCONFIG --no_trimming \ --species Hsap $SAMPLE $OUTPUT ${INPUT}_R1.fastq.gz ${INPUT}_R2.fastq.gz 5) Which input/output would be most helpful? A fastq or sam file?
Thank you, Alex
Hi Alex,
Fastq as input and the output directory for a couple of cells would be great. I will then try to run bracer and see if I can figure out what happens. Are you sure the cells are really B-lineage cells? Do you detect B-lineage markers and/or immunoglobulin genes?
Best, Ida
Ida,
These are B cells that our collaborator FACSed using (CD19+, CD3-).
Here is the link to a couple of fastq files and outputs: https://osu.box.com/s/dnlyvph7tucd6xiltjqbe91c3f5g1pl5
Thank you! Alex
Hi Alex, I have looked at the first two of the cells you sent me. I think there is an issue with the sorted cells - when looking at their entire gene expression they have expression of genes related to T cells (TCR genes, CD3, CD4 or CD8) for example, while no gene expression of B-cell releated genes. I would look at the gene expression of all your cells and identify if any of the sorted cells are in fact B cells.
Best, Ida
Bowtie2 seems to return very low VDJ alignment rates for each of my B cells and virtually no VDJ sequences were able to be reconstructed (BCR_summary.txt screenshot linked below).
Because bracer was built using plasma cells (15-20% IGHV mutation) from intestinal tissue, I am surprised that my 2 sets of data yields little to no BCR info (considering that each cell was sequenced 1-2 mil clusters). Could you provide us with your insight on why this might be happening?
*** Full-length transcriptome data from healthy donor B cells was generated using the Takara smart-seq HT kit (2x150bp). A screenshot of the sequencing depth and HISAT2 alignment to hg38 are linked below as well.
Thank you for your time and looking forward to hearing your insights, Alex