Segmentation fault (core dumped) during SVDSS search and SVDSS call with HiFi data

[comery]

Hello,

I am encountering a Segmentation fault (core dumped) error when running SVDSS search and SVDSS call on PacBio HiFi data. The data is aligned to the T2T-CHM13v2.0 reference genome.

The SVDSS index and SVDSS smooth steps completed successfully, but the pipeline crashes at the search step, resulting in an empty specifics.txt file. Consequently, the call step also fails with a segmentation fault after loading 0 SFSs.

Environment:

• Data type: PacBio HiFi (HG002)

• Reference: T2T-CHM13v2.0

• SVDSS version: [v2.1.0]

Commands executed:

# Set variables (paths anonymized)
REF='/path/to/GCF_009914755.1_T2T-CHM13v2.0_genomic_chr.fasta'
BAM='/path/to/hg002_to_chm13.hifi.sort.bam'

SVDSS index -t24 -d $REF > $REF.fmd 
SVDSS smooth --reference $REF --bam $BAM --threads 24 > smoothed.bam 
samtools index smoothed.bam 

The pipeline fails at these steps:

SVDSS search --index $REF.fmd --bam smoothed.bam > specifics.txt
SVDSS call --reference $REF --bam smoothed.bam --sfs specifics.txt --threads 24 --min-sv-length 50 > calls.vcf

Standard Output / Error Logs:

[2026-04-13 14:17:13.990] [stderr] [info] Loading reference genome from /path/to/GCF_009914755.1_T2T-CHM13v2.0_genomic_chr.fasta..
[2026-04-13 14:17:24.349] [stderr] [info] Max allowed alignment accuracy: 0.023337894416148718
[2026-04-13 14:17:24.566] [stderr] [info] Smoothing alignments on 24 threads..
Alignments processed so far: 21697113. Alignments processed per second: 9487. Time: 2287
[2026-04-13 14:55:31.837] [stderr] [info] All done! Runtime: 2298 seconds
smooth done
index done
[2026-04-13 15:06:37.555] [stderr] [info] Restoring index..
[2026-04-13 15:06:38.241] [stderr] [info] Extracting SFS strings on 4 threads..
Segmentation fault      (core dumped) SVDSS search --index $REF.fmd --bam smoothed.bam > specifics.txt

[2026-04-13 15:06:38.576] [stderr] [info] Loading reference genome from /path/to/GCF_009914755.1_T2T-CHM13v2.0_genomic_chr.fasta..
[2026-04-13 15:06:48.441] [stderr] [info] Loading SFSs from specifics.txt..
[2026-04-13 15:06:48.442] [stderr] [info] Loaded 0 SFSs from 0 reads.
[2026-04-13 15:06:48.506] [stderr] [info] Placing SFSs on reference genome
[2026-04-13 15:08:12.931] [stderr] [info] 0/0/0 unplaced SFSs. 0 erroneus SFSs. 0 clipped SFSs.
[2026-04-13 15:08:12.931] [stderr] [info] Clustering 0 SFSs..
Segmentation fault      (core dumped) SVDSS call --reference $REF --bam smoothed.bam --sfs specifics.txt --threads 24 --min-sv-length 50 > calls.vcf```
Could you please provide any insights on what might be causing this memory error or if there is a workaround for processing HiFi datasets against the T2T reference? Thank you for your time and assistance.

[ldenti]

Hello, this is odd. I have used CHM13 and HG002 HiFi successfully in the past. Not sure what's going on here...

How did you align the reads?

Could you share link to reads so I can try to reproduce the error?

Thanks,

[comery]

Hi,

The hifi datasets were downloaded from https://s3-us-west-2.amazonaws.com/human-pangenomics/index.html?prefix=T2T/scratch/HG002/sequencing/hifirevio/. I used the three of them (m84005_220827_014912_s1, m84039_230418_213342_s3, m84011_220902_175841_s1).

Read mapping was conduct following the method described in SVDSS paper as:
minimap verison: 2.26-r1175
samtools version: 1.19

minimap2  -t 12 -ax map-hifi --MD --eqx -Y  -R '@RG\tID:HG002' GCF_009914755.1_T2T-CHM13v2.0_genomic_chr.fasta m84011_220902_175841_s1.hifi_reads.fastq.gz|samtools view -bh -@ 12 -  > m84011_220902_175841_s1.hifi_reads.hifi.bam
samtools sort -@ 24 m84011_220902_175841_s1.hifi_reads.hifi.bam > m84011_220902_175841_s1.hifi_reads.hifi.sort.bam

And then merge three sorted bam together using samtools merge -@ 24 hg002_to_chm13.hifi.sort.bam *.hifi.sort.bam

[comery]

Hello, this is odd. I have used CHM13 and HG002 HiFi successfully in the past. Not sure what's going on here...

How did you align the reads?

Could you share link to reads so I can try to reproduce the error?

Thanks,

By the way, As you have successfully run the SVDSS in the past, could you share me with the SV results if you still have it?

[ldenti]

I ran SVDSS v2.1.1 without any issue. Here the VCF I got from the 3 FASTQ you used.

Here the calls I got from SVDSS using --min-cluster-weight 4: svdss2-minsupp4.vcf.gz

Here the commands:

for name in m84005_220827_014912_s1.hifi_reads m84011_220902_175841_s1.hifi_reads m84039_230418_213342_s3.hifi_reads.default
do
    minimap2  -t 32 -ax map-hifi --MD --eqx -Y -R '@RG\tID:HG002' \
              chm13v2.0.fa \
              $name.fastq.gz | samtools view -b | samtools sort -@31 > $name.bam
    samtools index $name.bam
done

samtools merge -@32 alignments.bam m*.bam
samtools index alignments.bam

SVDSS smooth --reference chm13v2.0.fa --bam alignments.bam --threads 32 > alignments-smoothed.bam
samtools index alignments-smoothed.bam

SVDSS search --threads 32 --index chm13.fa.fmd --bam alignments-smoothed.bam > specifics.txt

SVDSS call --threads 32 --reference chm13v2.0.fa --min-cluster-weight 4 --bam alignments-smoothed.bam --sfs specifics.txt > variants-minsupp4.vcf

(minimap v2.30-r1287 and samtools 1.19.2. And just for testing, I run smooth and search with SVDSS v2.1.0 and it worked).

Apart from tool versions, the other difference is the reference sequence (I used chm13v2.0.fa.gz). So, I am still not sure what was the issue...

I'll try to rerun using same versions and same reference sequence (can you please send me the URL to download it?)

Best,

[comery]

I appraciate you run this for me and thank a lot for sharing your results! It is very helpful!
Actually, the reference I used was T2T_CHM13 which was downloaded from NCBI (https://www.ncbi.nlm.nih.gov/datasets/genome/GCF_009914755.1/) and merely changed the sequence IDs to 'chrXXX' format and exclude the MT sequence. The faidx file looks like:

chr1	248387328	6	80	81
chr2	242696752	251492182	80	81
chr3	201105948	497222650	80	81
chr4	193574945	700842429	80	81
chr5	182045439	896837067	80	81
chr6	172126628	1081158080	80	81
chr7	160567428	1255436297	80	81
chr8	146259331	1418010824	80	81
chr9	150617247	1566098403	80	81
chr10	134758134	1718598373	80	81
chr11	135127769	1855040991	80	81
chr12	133324548	1991857865	80	81
chr13	113566686	2126848977	80	81
chr14	101161492	2241835254	80	81
chr15	99753195	2344261272	80	81
chr16	96330374	2445261389	80	81
chr17	84276897	2542795900	80	81
chr18	80542538	2628126266	80	81
chr19	61707364	2709675593	80	81
chr20	66210255	2772154307	80	81
chr21	45090682	2839192198	80	81
chr22	51324926	2884846521	80	81
chrX	154259566	2936813015	80	81
chrY	62460029	3093000832	80	81

Is that possible to cause this issue by a lower memery size? How large memery size when you run?