Single cell chemistry

Question

[David]

Hi all,

I'm a newbie in single cell and I a need to pre process some samples. Recently I did it with cellRanger count and it went fine with good results, but now my data was builded with a chemistry I have never seen before, I have 8 bp for I 25 bp for R1 and 90 bp for R2. The colleague who sent me the data says it was made with chemistry v2 but after after a few attempts it has been impossible to get the counts with cellranger count. I have read that cellranger needs at least 26 bp (barcode+UMI) in R1 to works properly.

Does anyone know how I should preprocess this data? and if it is possible to do it with cellranger?

Thanks

[GenoMax]

I did it with cellRanger count and it went fine with good results

It is not clear if you are referring to some other samples that worked well in past but you are now having trouble with a new set of samples. If you did not get any errors and the results look reasonable then you probably do not need to worry about chemistry.

Please post exact error message you are getting.

You can try to use alevin-fry or STARsolo as well.

[David]

Sorry, I did not explain it well.

I did it before with other group of samples and different chemistry, so cellranger is working fine in my system.

This is the error I received:

[error] Pipestance failed. Error log at:
11376PLpool10__Zfl2-2_1head_S9/SC_RNA_COUNTER_CS/SC_MULTI_CORE/MULTI_CHEMISTRY_DETECTOR/_GEM_WELL_CHEMISTRY_DETECTOR/DETECT_COUNT_CHEMISTRY/fork0/chnk0-u385a73807b/_errors

Log message:

The read lengths are incompatible with all the chemistries for Sample 11376PLpool10__Zfl2-2_1head in "/mnt/home/users/bio_028_genyo/twidmann/Transposable2/fastq/sample_11376PLpool10__Zfl2-2_1head_S9".
 - read1 median length = 25
 - read2 median length = 90
 - index1 median length = 8

The minimum read length for different chemistries are:

SFRP     - read1: 26, read2: 30, index1: 0
SC5P-R2  - read1: 26, read2: 25, index1: 0
SC5P-PE  - read1: 81, read2: 25, index1: 0
SC3Pv1   - read1: 25, read2: 10, index1: 14
SC3Pv2   - read1: 26, read2: 25, index1: 0
SC3Pv3   - read1: 26, read2: 25, index1: 0
SC3Pv3LT - read1: 26, read2: 25, index1: 0
SC3Pv3HT - read1: 26, read2: 25, index1: 0

We expect that at least 50% of the reads exceed the minimum length.

[GenoMax]

As you can see the data is one base short. If cell ranger is unable to analyze the data you may need to look elsewhere.

[swbarnes2]

OP could cheat a bit by adding an A to the end of every read of R1 (and a suitable letter to the quality string) and see if Cellranger can make anything remotely sensible from that.

[David]

I tried to add an "N" to the end of every read in R1 and cellranger did not found any barcode, results but no cells, I did it again adding a "N" to the start of each read, same result. Adding an "A" should be different?

[swbarnes2]

Well, A is a real letter. I don't know off the top of my head how cellranger deals with N's. But if the cell barcodes didn't match the v2 whitelist, then I'd say you don't have v2 chemistry.

Answer 1

[swbarnes2]

Are you sure these samples are 10x? How likey is it that the sequencing group messed up and ran the samples wrongly?

[David]

Unlikely, this data came from a experienced sequencing team, I 'm the inexperienced data analyst.

I'm not sure if these samples are 10x, I know these are single cell data and v2 chemistry according to my partner.

[swbarnes2]

As far as I know, v2 must mean 10x. But if it the cell barcode + umi is 26 bases, you don't have that.

[David]

That's the question, I think these data are not a useful entry for cellranger. Do you know if there is any other compatible software for them?

[swbarnes2]

You could still trick cellranger into processing them. Use umi tools to determine what the cell barcodes really are, (and see if they make sense) and temporarily replace the 10x provided whitelist with your whitelist. You might want to install the latest version of cellranger locally so you can do this without messing things up for others.

Or just stop and say that you aren't going to waste more time on data that looks messed up. V2 is supposed to have an R1 of 26 bases, and you don't have that, so stop playing detective, do nothing until you get definitive answers from someone as to what's going on.