Error when try to run ./rattle

[Kirovez]

Hi!
I try to run RATTLE with my fastq reads (direct RNA sequencing):
./rattle cluster -i my.fastq -t 24 --fastq --iso --rna -r 5

I got an error:

what(): cannot create std::vector larger than max_size()
RNA mode: 1 Reading fasta file... Done terminate called after throwing an instance of 'std::length_error' Aborted (core dumped)

Can you help with this, please?

Ilya

[novikk]

Hi Ilya,

how large is the file, and how much memory available do you have?

[Kirovez]

Hi,
I have 512Gb RAM and 1 Tb free disk space. My fastq has only 300,000 reads.

Ilya

[novikk]

And what is the size of the largest read? Is it possible for you to send me the fastq via email so that I can check what's going on?

Thanks

[Kirovez]

I tried to tun with different datasets and this error ('Segmentation fault (core dumped)') always occurred. The reads I used are publically available and can be downloaded via this link: https://sra-pub-src-2.s3.amazonaws.com/ERR3764351/col0_nanopore_drs_4.tar.gz.1 / I hope it will help in debugging.

[EduEyras]

Hi, thanks for your email. Have you tried processing the file to see whether it contains unusually long reads? sometimes chimeric reads are produced and need to be split. A massively long read may lead to those memory issues David Eccles and cols. wrote something about these cases: https://www.ncbi.nlm.nih.gov/pmc/articles/PMC5600009/ There are some things you could try: 1) You could check the distribution of read lengths and discard those that are clear outliers 2) Since you are probably detecting and removing adapters, you could try to identify internal adapters and split the reads on that point. I hope this helps Eduardo

…

On Thu, 13 Aug 2020 at 15:27, Ilya Kirov ***@***.***> wrote: I tried to tun with different datasets and this error ('Segmentation fault (core dumped)') always occurred. The reads I used are publically available and can be downloaded via this link: https://sra-pub-src-2.s3.amazonaws.com/ERR3764351/col0_nanopore_drs_4.tar.gz.1 / I hope it will help in debugging. — You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub <#17 (comment)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/ADCZKB5PTCD3AIPI6OIZDQ3SAN2V7ANCNFSM4PVNDMQQ> .

-- Prof. E Eyras EMBL Australia Group Leader The John Curtin School of Medical Research - Australian National University https://github.com/comprna http:https://scholar.google.com/citations?user=LiojlGoAAAAJ

[novikk]

Hi @Kirovez

I've checked the files you provided and I've found the problem. The issue is that these files were basecalled on windows, so there is an extra character before the line end (line ends in windows are \r\n while they are only \n in unix).

I will try to push a fix so that RATTLE accepts files basecalled in Windows ASAP, but meanwhile you can fix the line endings with the following command:

sed 's/\r//g' file.fq > file.fixed.fq
Tell me if that worked for you!