#3 [MUMMER/4.0.0beta2] nucmer4 --genome option and segmentation fault

[Fernando Cruz]

Dear Aleksey and Steven,

I have tried to run nucmer4 to compare a contiguous bird genome assembly against the chicken chromosomes (discarding the one with MHC in it). In order to do so I am just running nucmer with defaults and the --genome option and 24 threads (1TB). However I am getting an error saying segmentation fault:

cat test_nucmer4_module.log
remove PYTHON/2.7.5 (PATH)
remove jip-2.7/0.6 (PATH)
load GNUPLOT/5.2.7 (PATH)
load gcc/6.3.0 (PATH, MANPATH, LD_LIBRARY_PATH)
load PERL/5.22.1 (PATH, MANPATH, LD_LIBRARY_PATH)
load MUMMER/4.0.0beta2 (PATH)
running nucmer4 with genome option for large queries
./run_module_mummer4.0.0beta2.sh: line 29: 31318 Segmentation fault nucmer -t 24 --genome -p PavCri9_vs_Chicken_genome Gallus_gallus.chromNames.noChromMT.fa PavCri9.fa

real 15m2.108s
user 19m21.686s
sys 0m11.489s

running mummerplot with defaults on PavCri9_vs_Chicken_genome.delta

gnuplot 5.2 patchlevel 7
ERROR: Could not read PavCri9_vs_Chicken_genome.delta, File is empty

real 0m0.603s
user 0m0.062s
sys 0m0.026s

sed -n 29,29p evaluations/PavCri9/mummer4/run_module_mummer4.0.0beta2.sh
time nucmer -t 24 --genome -p PavCri9_vs_Chicken_genome Gallus_gallus.chromNames.noChromMT.fa PavCri9.fa

Note that, this is a fresh installation done by IT. The previpous one I installed myself and without the --genome option results were weird (single-dot alignments). With --genome option I´ve got same results as above.

Do you have any idea of how to fix this issue? Is there any new developmental version that we could test in our SLURM cluster? We are really insterested in assembling large genome assemblies for evaluation purposes or SV detection.

Thanks in advance,
Fernando