Re: EXITING because of FATAL ERROR: Read1 and Read2 are not consistent

[Babak A]

Hello,

I know why that happens as I faced the exact same problem last week! It is because read_1.fastq and read_2.fastq have different number of lines (reads). I surprisingly found out that after dumping an SRA file, one of the files was 4 GB smaller than the other mate's file.

Use wc -l read_*.fastq and if they have different number of lines contact SRA.

Cheers

Babak

On Saturday, June 1, 2013 5:42:35 PM UTC-4, Johnbosco Tayebwa wrote:

I think this error has not been reported before. I am using STAR 2.3.1n on Ubuntu 12.04.1 and I keep getting the error

"EXITING because of FATAL ERROR: Read1 and Read2 are not consistent reached the end of the one before the other one
SOLUTION: Check you your input files: they may be corrupted"

after which STAR exits. The command I am using is:

./STAR --genomeDir /path/to/genomeDir --readFilesIn Read1 Read2 --readFilesCommand zcat --outFileNamePrefix  sampleID_star --chimSegmentMin 20 --runThreadN 24

This happens for some samples while for others, STAR works fine.

I would really appreciate a quick response since I need results urgently. Thanks in advance!

[Babak A]

And you might want to use --split-3 with fastq-dump. It is mentioned to be a legacy option, but SRA help desk actually suggests using it.

Babak

[Johnbosco Tayebwa]

Thanks Babak for your explanation. My reads do not come from SRA. They were obtained from a paired-end sequencing run from Illumina's HiscanSQ using the TruSeq RNA Sample Prep Kit v2. I will use wc -l to check the number of lines and get back to you.

[Shawn Driscoll]

If the problem is that one of the two mate files is missing reads relative to the other the solution is kind of annoying but possible.

[Johnbosco Tayebwa]

Sample that gave an error: Read1 5192019
                                            Read2 5221858

Sample that didn't give an error: Read1 5302511
                                                    Read2 5348421

So it doesn't seem like the different number of lines should be the problem here. Shawn, what would that solution be?

[Alexander Dobin]

Hi Johnbosco,

both samples should have given the error, since the number of lines is different for mates 1,2. 

Are you doing:

zcat Read{1,2}.gz | wc -l

If you could send me these samples, I will have a closer look at them.

Cheers

Alex

[Johnbosco Tayebwa]

Dear Alex,

Sorry, I had used "wc -l" forgetting the files had  *.gz extension. using "zcat Read{1,2}.gz | wc -l", all the files have the same number of lines. I will send you one sample giving the above error using my gmail account. Please let me know when you have received them.

[Shawn Driscoll]

I've heard of people getting reads back from their sequencing people and the two files aren't sorted correctly which makes the aligners flip out a little.  The best correction I've seen suggested is to convert both FASTQ files to SAM format, merge them, sort by read name and then split them back out to FASTQ format. I don't know if theres a tool that does all of those steps for you though so it would require a bit of tinkering.  Also if you really had different numbers of reads it would be kind of a messy process.  I've never had to deal with that myself so I don't really have any better suggestions.  I'm sure Alex will figure this one out.

[Alexander Dobin]

Hi  Johnbosco,

I mapped your files and it was completed without any problems, so the files appear to be intact. I am thinking that this error could be caused by a premature termination of one of the zcat processes. Could you please try to re-run this example to check this? If it does not work, and you need the results quickly, please try to unzip the files and map them without  --readFilesCommand zcat. Later we can try to debug the problem, which could be system dependent.

Cheers

Alex

[Johnbosco Tayebwa]

Alright. Thanks. Will do that and get back to you.

[Satshil Rana]

I'm having the same problem as you are.

My run consists of 4 lanes. STAR gets through lanes 1-3 without error, but when it starts on lane4, I get the same exact error as you "EXITING because of FATAL ERROR: Read1 and Read2 are not consistent, reached the end of the one before the other one

SOLUTION: Check you your input files: they may be corrupted"

I just did a wc -l on lane4 and they're both the same size/length.

BWA was able to properly align this read set, so far I've only experienced this issue in STAR.

Is there something else I can do to fix the problem?

[Alexander Dobin]

Hi Satshil,

please send me the Log.out of this run.

Cheers

Alex

[Satshil Rana]

Hey Alex,

The weird thing about the run: If I ran it with 4 lanes together then it crashes giving me that error on lane 4. If I run lane 4 by itself then it successfully finishes the mapping. It's pretty weird. I deleted the log of when it crashed since I was able to run the 4 separate lanes and merge them together, but I can re-run it and give you the log when it crashes if you want me to do that.

Thanks,

Satshil

[Alexander Dobin]

Hi Satshil,

it could be something like an empty line at the end of lane 3.

Then each of the lanes would work, but all together they will create invalid stream.

Cheers

Alex