https://github.com/fertiglab/sra-tools-notes

Some notes on sra-tools usage

https://github.com/fertiglab/sra-tools-notes

Repository

Some notes on sra-tools usage

https://github.com/FertigLab/sra-tools-notes/blob/master/

# Some notes on sra-tools usage
It is an outline of sra-tools usage routine. First, the copmputer (e.g. AWS instance) is to have enough memory (16G? -- maybe, 8G), and to have disk space to keep the raw data for at least one experiment. It could be 250Gish if we plan to work with RNASeq fastq/fasta.

## Installation
Sra-tools as well the aligner or other tools are very convienient to be installed with conda. First, install conda itself. 
The [conda installation instructions](https://docs.conda.io/projects/conda/en/latest/user-guide/install/linux.html) explains what to do with the [conda installation script](https://docs.conda.io/en/latest/miniconda.html "Miniconda installtion script"). To download the script using the command-line, you can use `wget` command in the shell. 

Conda suggest to modify the shell prompt after installation. I wuold say no and just add conda/bin (usually, it is ~/mniconda3/bin) to path. Now, install [sra-tools](https://anaconda.org/bioconda/sra-tools "sra-tooks installtion page in conda"), [salmon](https://anaconda.org/bioconda/salmon "salmon installtion page in conda"). etc. You are to do it once for a new user, then sometimes [update conda](https://docs.conda.io/projects/conda/en/latest/commands/update.html) if you like.

[Here are the detailed instruction how to install conda, sra-tools and salmon on aws instance](https://github.com/FertigLab/sra-tools-notes/blob/master/conda-on-aws-stepwise.md).

## dbGaP key and folder
Usually, the FastQ file is controlled thing and you are to have access. Technically, it means that you have the keyfile from dbGaP. Usually, its name is something like prj_NNNNNNN.ngc. NNNNN is a number of dbGaP -- our is 23539. To start working with the project, say:

`vdb-config --import prj_23539.ngc`

The command will import the keyfile and it will create the project work folder ~/ncbi/dbGaP-23539 (project number is the dbGaP projret number you work with). It happens only once. All the commands like prefetch, fatq-dump, etc are to be run form this folder. You are not supposed to change the folder name.

## dbGaP accession number

To download a dbGaP file using sra-tools, you are to know the file's SRR accesion ID. The [Run Selector](https://trace.ncbi.nlm.nih.gov/) website shows the IDs. Go to dbGaP page of the study, e.g. (https://www.ncbi.nlm.nih.gov/projects/gap/cgi-bin/study.cgi?study_id=phs001709.v1.p1), and in the left lower corner of 'Molecular datasets' tab click the 'Run selector' link. Choose the data type (e.g., 'RNA'), etc. Select all the files you need, and then download the text file with the SRR IDs of all the files you need (accesion list). Alternatively, you can download RuniInfoTable that contains a text table, a line per SRR, with SRR's as first field and a lot of other information. 

## FastQ 
Usually, we do not need all the fatstq files simultaneously -- so, the best option is to write a script that reads the SRR list, and for each SRR does the foolowing (again, the working folder for all sra-tools is: ~/ncbi/dbGaP-23539).

`cd ~/ncbi/dbGaP-23539`

`prefetch SRR10003688`

`fastq-dump --split-files --outdir  sra/SRR10003688.sra`

Now, go to my-fastq-path and quantify the fatq files. Then, remove them and ~/ncbi/dbGaP-23539/sra/SRR10003688.sra. Here, SRR10003688 is the first SRR from the phs001709.v1.p1 project and it is a templeate for SRR.

Good luck!

Owner

GitHub Events