[nupurgupta]

Same problem

Did you find a solution to the null problem please?

Hi everybody,

I am have similar problems with samtools 0.1.18. I would like to have reference characters listed in a pileup files, but I have problems with headers.

samtools faidx AGSbrut.fasta
samtools view -q 20 -buh -t AGSbrut.fasta.fai A.sam | samtools sort - A
samtools view -q 20 -buh -t AGSbrut.fasta.fai S.sam | samtools sort - S
samtools mpileup -B -f AGSbrut_index.fai A.bam S.bam > AS.mpileup

[fai_build_core] different line length in sequence 'null'.
Segmentation fault

I hypothesized that this 'null' sequence may be a blank line; so I looked for it manually and with sed, with no luck. I also looked for other potential problems based on what was previously reported (no extra spaces, characters, etc in reference sequence names in fai and sam files). I also tried to re-head the file, with no success:

samtools view -HS -t AGSbrut.fasta.fai A.sam > Aheader.sam
samtools reheader Aheader.sam A.bam > Aheaded.bam

[bam_header_read] EOF marker is absent. The input is probably truncated.

All insights are welcome!
thank you, eric

[jgibbons1]

Using pileup with the -f argument allows you to supply the faidx indexed reference sequence file. I used this option and it fixed my problem.

[jgibbons1]

Hey folks,

Have been struggling to figure out why I am getting N's for my pileup reference sequence. I found hope when I discovered this string but I have followed all the suggestions to no avail. I've tried this with different versions of samtools, different data sets, different reference files and have simplified ID names, rebuilt the faidx index, etc. etc.

Still can't figure out what's going on here. Has anyone found any other solutions?

Thanks

[adowney]

Here's another possible solution - the headers are not consistent between SAM/BAM and the original fasta:

Even though the reference file was the same one in both cases, sometimes aligners just write a substring out into the SAM file. Samtools seems to take the full header.

For example the first contiguous part of my genome header is
gi|110645304|ref|NC_002516.2|

However in my SAM file the aligner has only written
NC_002516.2

Samtools has written the full header to the .fa.fai index
gi|110645304|ref|NC_002516.2|

.. and this does not match.

Solution:

Try correcting the original header on the reference fasta to just the substring which the aligner uses.
eg
gi|110645304|ref|NC_002516.2|
to
NC_002516.2

The above suggestion fixed the problem when I got this error

[ericpante]

Hi everybody,

I am have similar problems with samtools 0.1.18. I would like to have reference characters listed in a pileup files, but I have problems with headers.

samtools faidx AGSbrut.fasta
samtools view -q 20 -buh -t AGSbrut.fasta.fai A.sam | samtools sort - A
samtools view -q 20 -buh -t AGSbrut.fasta.fai S.sam | samtools sort - S
samtools mpileup -B -f AGSbrut_index.fai A.bam S.bam > AS.mpileup

[fai_build_core] different line length in sequence 'null'.
Segmentation fault

I hypothesized that this 'null' sequence may be a blank line; so I looked for it manually and with sed, with no luck. I also looked for other potential problems based on what was previously reported (no extra spaces, characters, etc in reference sequence names in fai and sam files). I also tried to re-head the file, with no success:

samtools view -HS -t AGSbrut.fasta.fai A.sam > Aheader.sam
samtools reheader Aheader.sam A.bam > Aheaded.bam

[bam_header_read] EOF marker is absent. The input is probably truncated.

All insights are welcome!
thank you, eric

[smehr12]

I had this same problem, and after seeing no solution here did some more digging, and have a possible solution for you.

I noticed that the ref.fa.fai file for my whole genome was 0 kb. The .fai is used by samtools when building the pileup. When I ran the command to re-build the .fai:

samtools faidx reference.fa

I got the following error message:

[fai_build_core] different line length in sequence 'scaffold_14'.
Segmentation fault

No doubt this same message occurred the first time I ran the pileup command (which also builds the .fai if it doesn't exist), but I apparently didn't pay attention. After that first time, the .fai file EXISTED so no errors were subsequently reported when I ran pileup again.

In my case, there was an extra line after scaffold_14. I removed this, and re-built the .fai using the samtools faidx command and then re-ran the pileup command. My pileup then contained the reference base as intended!

Hope this helps y'all find the solution to your problem.
Best,
Shannon
University of Texas at Austin

Hi all,
I have the same error.
samtools faidx bwa.ref/ref.fasta ref.fa

ERROR:
different line length in sequence 'scaffold_67'.
Segmentation fault
NOTE: I see NNNN in that scaffold . Does anyone have a suggestion?

[bgibb]

I noticed the same problem when running pileup under SAMtools-0.1.15. However the problem does not seem to occur when running pileup under SAMtools-0.1.4 (using the same reference file, same BAM file and same command line options).

samtools-0.1.4/samtools pileup -s -f reference.fa sorted.bam > pileup.out

[colindaven]

Here's another possible solution - the headers are not consistent between SAM/BAM and the original fasta:

Even though the reference file was the same one in both cases, sometimes aligners just write a substring out into the SAM file. Samtools seems to take the full header.

For example the first contiguous part of my genome header is
gi|110645304|ref|NC_002516.2|

However in my SAM file the aligner has only written
NC_002516.2

Samtools has written the full header to the .fa.fai index
gi|110645304|ref|NC_002516.2|

.. and this does not match.

Solution:

Try correcting the original header on the reference fasta to just the substring which the aligner uses.
eg
gi|110645304|ref|NC_002516.2|
to
NC_002516.2

[smol]

Hi
I'm having the same problem with Ns in my pileup file and have tried everything mentioned above (thanks for suggestions!). I am using:

./samtools pileup data.sorted.bam -f reference.fasta > data.pileup

My reference .fai file looks like this:

chr2L 49364325 7 60 61
chr2R 61545105 50187078 60 61
chr3L 41963435 112757942 60 61
chr3R 53200684 155420775 60 61
chrUNKN 42389979 209508147 60 61
chrX 24393108 252604632 60 61
chrY 237045 277404298 60 61

Any ideas?

[brutus]

I also had this experience, in my case the problem disappeared when I removed spaces in the reference sequence name.

[mmartin]

I had the same problem. In my case, I had colons in the reference sequence names, something like "Region1:1-100". When I removed them, samtools pileup worked as expected.

[hollandorange]

I got the same problem.
chr17 418628 N 54
chr17 418629 N 58
chr17 418630 N 57

[skingan]

It turned out to be a similar problem to the one SMHfrog had. In my reference file, each chromosome sequence was on a single line, so when samtools built the .fai file there was a segmentation fault because of the length of the sequence. I used a different reference with line breaks and it worked. I used the same reference file for the Mosaik run and the pileup build.
Sarah

[thaley]

Ran into the same problem. It may be worth someone adding this to the faidx documentation regarding null strings in the reference or make the thrown error more descriptive.

[SMHfrog]

I had this same problem, and after seeing no solution here did some more digging, and have a possible solution for you.

I noticed that the ref.fa.fai file for my whole genome was 0 kb. The .fai is used by samtools when building the pileup. When I ran the command to re-build the .fai:

samtools faidx reference.fa

I got the following error message:

[fai_build_core] different line length in sequence 'scaffold_14'.
Segmentation fault

No doubt this same message occurred the first time I ran the pileup command (which also builds the .fai if it doesn't exist), but I apparently didn't pay attention. After that first time, the .fai file EXISTED so no errors were subsequently reported when I ran pileup again.

In my case, there was an extra line after scaffold_14. I removed this, and re-built the .fai using the samtools faidx command and then re-ran the pileup command. My pileup then contained the reference base as intended!

Hope this helps y'all find the solution to your problem.
Best,
Shannon
University of Texas at Austin