Tag Archives: bash

Data Management – Integrity Check of Final BGI Olympia Oyster & Geoduck Data

After completing the downloads of these files from BGI, I needed to verify that the downloaded copies matched the originals. Below is a Jupyter Notebook detailing how I verified file integrity via MD5 checksums. It also highlights the importance of doing this check when working with large sequencing files (or, just large files in general), as a few of them had mis-matching MD5 checksums!

Although the notebook is embedded below, it might be easier viewing via the notebook link (hosted on GitHub).

At the end of the day, I had to re-download some files, but all the MD5 checksums match and these data are ready for analysis:

Final Ostrea lurida genome files

Final Panopea generosa genome files

Jupyter Notebook: 20161214_docker_BGI_data_integrity_check.ipynb

Data Analysis – Identification of duplicate files on Eagle

Recently, we’ve been bumping into our storage limit on Eagle (our Synology DS413):


Being fairly certain that there’s a significant amount of large datasets that is duplicated throughout Eagle, I ran a program on Linux called “fslint”. It searches for duplicates files based on a few parameters and is smart enough to be able to compare files with different filenames that share the same file contents!

I decided to check for duplicate files in the Eagle/archive folder and the Eagle/web folder. Initially, I tried searching for duplicates across all of Eagle, but after a week of running I got tired of waiting for results and ran the analysis on those two directories independently. As such, there is a possibility that there are more duplicates (consuming even more space) across the remainder of Eagle that have not been identified. However, this is a good starting point.

Here are the two output files from the fslint analysis:


To get a summary of the fslint output, I tallied the total amount of duplicates files that were >100MB in size. This was performed in a Jupyter notebook (see below):
Notebook Viewer: 20160114_wasted_space_synologies.ipynb
Jupyter (IPython) Notebook File: 20160114_wasted_space_synologies.ipynb


Here are the cleaned output files from the fslint analysis:



There are duplicates of files (>100MB in size) that are consuming at least 730GB!

Since the majority of these files exist in the Eagle/web folder, careful consideration will have to be taken in determining which duplicates (if any) can be deleted since it’s highly possible that there are notebooks that link to some of the files. Regardless, this analysis shows just how space is being consumed by the presence of large, duplicate files; something to consider for future data handling/storage/analysis with Eagle.

Data Received – Oly 2bRAD Illumina Sequencing from Genewiz

The data was made available to use on 20151224 and took two days to download.

The full list of samples (and the individual samples/libraries/indexes) submitted to Genewiz for this project by Katherine Silliman & me can be seen here (Google Sheet): White_BS1511196_R2_barcodes

The data supplied were all of the Illumina output files (currently not entirely sure where/how we want to store all of this, but we’ll probably want to use them for attempting our own demultiplexing since there were a significant amount of reads that Genewiz was unable to demultiplex), in addition to demultiplexed FASTQ files. The FASTQ files were buried in inconvenient locations, and there are over 300 of them, so I used the power of the command line to find them and copy them to a single location: http://owl.fish.washington.edu/nightingales/O_lurida/2bRAD_Dec2015/

Find and copy all FASTQ files:

find /run/user/1000/gvfs/smb-share\:server\=owl.fish.washington.edu\,share\=home/ -name '*.fastq.*' -exec cp -n '{}' /run/user/1000/gvfs/smb-share\:server\=owl.fish.washington.edu\,share\=web/nightingales/O_lurida/ \;

Code explanation:

  • Command line program used for searching for files
  • Location of the files I wanted to search through. The path looks a little crazy because I was working remotely and had the server share mounted.
-name '*.fastq.*'
  • The name argument tells the find command to look for filenames that have “.fastq” in them.
-exec cp -n '{}'
  • The exec option tells the find command to execute a subsequent action upon finding a match. In this case, I’m using the copy command (cp) and telling the program not to overwrite (clobber, -n) any duplicate files.
/run/user/1000/gvfs/smb-share\:server\=owl.fish.washington.edu\,share\=web/nightingales/O_lurida/2bRAD_Dec2015 \;
  • The location where I want the matched files copied.

I created a readme file in the directory directory with these files: readme.md

I wanted to add some information about the project to the readme file, like total number of sequencing reads generated and the number of reads in each FASTQ file.

Here’s how to count the total of all reads generated in this project

totalreads=0; for i in *.gz; do linecount=`gunzip -c "$i" | wc -l`; readcount=$((linecount/4)); totalreads=$((readcount+totalreads)); done; echo $totalreads

Total Reads: 588,396,334

Code explanation:

  • Creates variable called “totalreads” and initializes value to 0.
for i in *.gz;
  • Initiates a for loop to process any filenames that end with “.gz”. The FASTQ files have been compressed with gzip and end with the .gz extension.
do linecount=
  • Creates variable called “linecount” that stores the results of the following command:
`gunzip -c "$i" | wc -l`;
  • Unzips the files ($i) to stdout (-c) instead of actually uncompressing them. This is piped to the word count command, with the line flag (wc -l) to count the number of lines in the files.
  • Divides the value stored in linecount by 4. This is because an entry for a single Illumina read comprises four lines. This value is stored in the “readcount” variable.
  • Adds the readcount for the current file and adds the value to totalreads.
  • End the for loop.
echo $totalreads
  • Prints the value of totalreads to the screen.

Next, I wanted to generate list of the FASTQ files and corresponding read counts, and append this information to the readme file.

for i in *.gz; do linecount=`gunzip -c "$i" | wc -l`; readcount=$(($linecount/4)); printf "%s\t%s\n%s\t\t\n" "$i" "$readcount" >> readme.md; done

Code explanation:

for i in *.gz; do linecount=`gunzip -c "$i" | wc -l`; readcount=$(($linecount/4));
  • Same for loop as above that calculates the number of reads in each FASTQ file.
printf "%s\t%s\n\n" "$i" "$readcount" >> readme.md;
  • This formats the the printed output. The “%s\t%s\n\n” portion prints the value in $i as a string (%s), followed by a tab (\t), followed by the value in $readcount as a string (%s), followed by two consecutive newlines (\n\n) to provide an empty line between the entries. See the readme file linked above to see how the output looks.
>> readme.md; done
  • This appends the result from each loop to the readme.md file and ends the for loop (done).