Step 4: error in Parsing SE reads

[Anoop Anand Malik]

Dear Aurthur,

Hope you are doing well!!

I am using usearch instead of vsearch to develop mock reference using the following command in step 4

perl /home/sbt/Documents/GBS-SNP-CROP-Ke/parsed/demultiplexed/GBS-SNP-CROP-4.pl -d SE -b ke_barcodesID.txt -rl 150 -pl 32 -p 0.01 -id 0.93 -t 10 -MR MockRef_ke


The same command worked for my previous analysis few months ago but this time I am getting the following error repeatedly:

#################################
# GBS-SNP-CROP, Step 4, v.2.0
#################################
Parsing Single-End reads...
GBS-SNP-CROP cannot access the sizes of the files selected to build the Mock Reference and so cannot calculate the proportion of those files to be used to randomly sample your data.
Please refer to the GBS-SNP-CROP User Manual 'STAGE2: BUILD THE MOCK REFERENCE' for more details.

I would kindly request you to resolve this problem. Also please find attached the script and barcode file (tab delimited) which I am using for this analysis. All the files which I got from step 3 are of 8 to 40 MB's, so none of them is 0 bytes.

Hope to hear from you soon.

Regards

Dr Anoop Anand Malik
India

[Arthur Melo]

Hi Annop,

Are you running the script in the same folder (directory) the files you selected to build de mock reference is located?

[Anoop Anand Malik]

Dear Arthur,

Yes, I am running the script from demultiplexed folder. The previous script (step 3) I ran from "parsed" folder, then a new folder was created "demultiplexed". In demultiplxed folder I obtained 96 files with a name "Lib1_01.fq.gz, Lib1_02.fq.gz .... and so on till Lib1_96.fq.gz". Then I copied my barcodesID file from parsed folder to demultiplxed folder and selectively changed YES to NO for constructing the mock reference. Here I was using the old script where usearch commands were present instead of new script where you have modified it to vsearch. In the script 4, I also changed the destination of usearch.

I used the same method last year (October 2019) to develop the mock reference and faced the similar problem.

I would also like to request you to guide me about usage and command line to install vsearch as from next time onward I would like to use vsearch instead usearch. I downloaded the latest version of vsearch. In the vsearch folder I ran following commands ./configure, make and make install, the make install command is not working as it is showing error 1. Hence the the vsearchIN.fa file is of 0 bytes. I would request you to guide me about this as well.

Regards

Dr Anoop Anand Malik

India  

[Arthur Melo]

Hi Anoop,

If you want to use vsearch instead usearch, please, use the latest version of GBS-SNP-CROP...

As you said your files have the extension *.fq.gz, I supposed you're using scripts from different pipeline versions. The v.2.0 does not compress the outputs of any step and GBS-SNP-CROP.4.pl is looking for *.fastq files instead *.fq.gz and once the script is not finding the files it can not calculate them sizes...

Once you are parsing single-end files, please go to GBS-SNP-CROP.4.pl script and make some changes:

line 416:

my $R1input1 = join (".", "$file","R1","fastq"); --> my $R1input1 = join (".", "$file","fq","gz");

line 418:

open my $IN, "<", "$R1input1" or die "Can't open $R1input1\n"; --> open my $IN, '-|', 'gzip', '-dc', $R1input1 or die "Can't open file $R1input1: $!\n";

If you have trouble to install vsearch or usearch, I strongly recommend you ask for help with the pipeline developers.

Please, let me know how it goes.

Best,

Arthur