No alignment with Brunello library file

[Johannes Schlondorff]

Dear Jan,

First, thank you for developing such a wonderful pipeline for CRISPR data analysis for bio-informatic novices like myself.

I do seem to be running into an issue, however.  I attempted to align sequence data from a CRISPR screen with the Brunello library file that I downloaded from CRISPR-analyzer.  While I achieved reasonable sgRNA extraction rates (~85%) and quality scores were reasonable, less than 2% of the sgRNA reads could be mapped to the Brunello library file!  I have manually gone and picked out a few random sgRNA sequences from my fasta files and match them to Brunello library sequences, and get much higher hit rates, so I don't think it is an issue with having been given the incorrect CRISPR library.  Any thoughts for whether this might be an issue with how I set up the analysis?

Best,

Johannes Schlondorff

[Jan Winter]

Dear Johannes,

thank you for your post.

Unfortunately I do not have raw sequencing data available to test the Brunello library, but there a couple of things that might have gone wrong.

In order to elucidate the error, could you please do the following:

- did you download the FASTA file from within the CRISPR-Analyzer (help section in Data Upload) or from the Github directly?

- In case you used the web-service, please repeat what you did and then go to help -> send me a ticket.

- In case you run it using docker, please redo the steps and send me screenshot of the Data Review tab

In case you have used the docker version, please proceed with the following in addition to the screenshot from above:

- please get the latest beta version by `docker pull boutroslab/crispranalyzer:1.16BETA` and run this (in Kitematic, remove the docker image and download the beta version by clicking on the TAGS field in Kitematic)

- redo all the steps you did (FASTQ extraction)

Then please email me the following:

- send me a screenshot of the data review tab (so one screenshot before the beta and one with the beta, since I have changed a couple of things)

- go to About CRISPRAnalyzeR and click on Download Log files -> please send me this as well

You can send me everything to jan.w...@dkfz-heidelberg.de -  I will keep this confidential.

As the sgRNA extraction itself goes fine, it is weird that bowtie2 mapping reveals low mapping rates.

One things that might have occurred is that the FASTQ extraction regular expression that can be selected in Step3 was wrong. But I guess you selected the default one for the backbone, right?

Maybe I have an issue there regarding the backbone, I will also check it.

Sorry for the troubles you experience, But I am sure we get this solved :)

Thanks for your help!

Jan

[Johannes Schlondorff]

Dear Jan,

Thanks for looking into this.

I think I may have figured out the issue.  I repeated the analysis, but after looking back at the sequence data and guide data, I think that the default FASTA extraction for the Brunello (pLentiGuide v2) backbone is wrong.  As listed, it is ACC(etc), but I believe it should be ACCG (which I think is one of the defaults for another backbone).  When I changed this parameter manually, I got >90% mapping and much better results.  (I can't send you screen shots right now; I kept getting booted off the server and my organizations IT doesn't let me turn on the virtualization feature on my work computer, so I haven't been able to download and run the application locally.)

Anyway, thanks again.  I'll let you know how things progress with further analysis.

Hanno