Numbat running for two samples separately but crashes when I merge the two and perform the same analysis again

[quail768]

Hey Teng,

Really amazing tool!
I am trying to run Numbat on two samples from the same patient.

I have run the Pileup and Phase script for both samples individually and Numbat seems to work when I run them individually(2 separate Numbat runs for 2 separate samples)

When I merge the two samples and try to run Numbat on the merged sample, I get the following error:

INFO [2024-08-26 11:27:34] 
numbat version: 1.4.1
scistreer version: 1.2.0
hahmmr version: 1.0.0
Running under parameters:
t = 1e-07
alpha = 1e-04
gamma = 20
min_cells = 40
init_k = 10
max_cost = 12468.9
n_cut = 6
max_iter = 2
max_nni = 100
min_depth = 0
use_loh = auto
segs_loh = None
call_clonal_loh = FALSE
segs_consensus_fix = Given
multi_allelic = TRUE
min_LLR = 5
min_overlap = 0.45
max_entropy = 0.7
skip_nj = FALSE
diploid_chroms = None
ncores = 15
ncores_nni = 15
common_diploid = TRUE
tau = 0.3
check_convergence = TRUE
plot = TRUEa
genome = hg38
Input metrics:
41563 cells
INFO [2024-08-26 11:27:35] Mem used: 10.8Gb
INFO [2024-08-26 11:28:06] Approximating initial clusters using smoothed expression ..
INFO [2024-08-26 11:28:08] Mem used: 10.8Gb
INFO [2024-08-26 13:51:18] running hclust...
INFO [2024-08-26 14:05:29] Iteration 1
INFO [2024-08-26 14:05:33] Mem used: 35.7Gb
INFO [2024-08-26 14:06:03] Expression noise level (MSE): medium (1.3). 
INFO [2024-08-26 14:06:03] Using fixed consensus CNVs
INFO [2024-08-26 14:07:56] Running HMMs on 10 cell groups..
ERROR [2024-08-26 14:08:15] No segments containing more than 10 genes for CHROM 21.
ERROR [2024-08-26 14:08:15] Error in smooth_segs(., min_genes = min_genes) : 
No segments containing more than 10 genes for CHROM 21.

I have re-ran the Pileup and Phase for both samples together as instructed in the manual.
Numbat runs for the same samples separately so I'm assuming that there are in-fact, genes for CHROM21

Any thoughts on what could be going wrong or something I could change?

Thanks for the help!

[teng-gao]

You should run pileup and phase jointly for both samples. Please check the command line options for pileup_and_phase.R for how to do this. Thanks

[quail768]

Hey @teng-gao .
Thanks for the reply.
Sorry if I was unclear.

Like I said earlier, I did run pileup and phase jointly for both samples that I planed to use for the combined Numbat run. From that single pileup and phase run, I have multiple allele_df files, all of which were included in the main Numbat run. I have run the multi-sample mode before with no problems.

Any thoughts?

[boneoncology]

I encountered the same issue, there is nothing wrong with seperate sample running, but error came out when I merge them together. I followed the guideline to make unique prefixed-sample barcode for each reads in merged BAM to run pile up phase.

[teng-gao]

Hi @boneoncology could you share some example data (expression matrix and allele df) to reproduce the error? Doesn't have to be the full dataset. Thanks

[boneoncology]

Hi @boneoncology could you share some example data (expression matrix and allele df) to reproduce the error? Doesn't have to be the full dataset. Thanks

out = run_numbat(

• count_mat_os, # gene x cell integer UMI count matrix
• ref_hca,
• df_allele_filtered,
• genome = "hg38",
• t = 1e-5,
• ncores = 32,
• plot = TRUE,
• out_dir = "/home/MyPassport/sda/numbat/OS_CELL",min_genes = 1)

numbat version: 1.4.2
scistreer version: 1.2.0
hahmmr version: 1.0.0
Running under parameters:
t = 1e-05
alpha = 1e-04
gamma = 20
min_cells = 50
init_k = 3
max_cost = 5841.6
n_cut = 0
max_iter = 2
max_nni = 100
min_depth = 0
use_loh = auto
segs_loh = None
call_clonal_loh = FALSE
segs_consensus_fix = None
multi_allelic = TRUE
min_LLR = 5
min_overlap = 0.45
max_entropy = 0.5
skip_nj = FALSE
diploid_chroms = None
ncores = 32
ncores_nni = 32
common_diploid = TRUE
tau = 0.3
check_convergence = FALSE
plot = TRUE
genome = hg38
Input metrics:
19472 cells
Mem used: 26.6Gb
Approximating initial clusters using smoothed expression ..
Mem used: 26.6Gb
number of genes left: 11109
running hclust...

Iteration 1
Mem used: 26.6Gb
Expression noise level (MSE): high (1.9). Consider using a custom expression reference profile.
Running HMMs on 5 cell groups..
错误于find_common_diploid(bulks, gamma = gamma, alpha = alpha, ncores = ncores):
Error in smooth_segs(., min_genes = min_genes) :
No segments containing more than 10 genes for CHROM 18.
此外: 警告信息:
1: In asMethod(object) :
sparse->dense coercion: allocating vector of size 2.0 GiB
2: In asMethod(object) :
sparse->dense coercion: allocating vector of size 1.6 GiB
3: In mclapply(bulks %>% split(.$sample), mc.cores = ncores, function(bulk) { :
预定内核5, 2的用户代码中出了错，所有作业值都将受到影响

count_mat_os[1:10,1:10]
BS1_AAACCCACAATGAAAC-1 BS1_AAACCCAGTGCACATT-1 BS1_AAACCCATCCGCTGTT-1 BS1_AAACCCATCTCATAGG-1 BS1_AAACCCATCTTTGGAG-1 BS1_AAACGAAAGGGCTTCC-1 BS1_AAACGAACACAGTCAT-1 BS1_AAACGAACACCTTCCA-1
AL627309.1 0 0 0 0 0 0 0 0
AL627309.5 0 0 0 0 0 0 0 0
AP006222.2 0 0 0 0 0 0 0 0
AL669831.2 0 0 0 0 0 0 0 0
LINC01409 0 0 1 0 0 1 0 1
FAM87B 0 0 0 0 0 0 0 0
LINC01128 0 0 0 1 1 0 0 0
LINC00115 0 0 0 0 0 0 0 1
FAM41C 0 0 0 0 0 0 0 0
AL645608.6 1 0 0 1 0 0 0 0
BS1_AAACGAACACGATTCA-1 BS1_AAACGAAGTGTTTACG-1
AL627309.1 0 0
AL627309.5 0 0
AP006222.2 0 0
AL669831.2 0 0
LINC01409 1 0
FAM87B 0 0
LINC01128 0 0
LINC00115 0 1
FAM41C 0 0
AL645608.6 1 1
head(df_allele_filtered)
cell snp_id CHROM POS cM REF ALT AD DP GT gene
1 BS3_GACGTTAGTGACTGTT-1 1_263722_C_G 1 263722 0 C G 0 1 1|0
4 BS12_AACACACCAAACCACT-1 1_779047_G_A 1 779047 0 G A 1 1 0|1
5 BS12_AACCCAACACCTTCGT-1 1_779047_G_A 1 779047 0 G A 0 1 0|1
7 BS12_ACCAAACCACTATCCC-1 1_779047_G_A 1 779047 0 G A 1 1 0|1
10 BS12_CAACCTCGTGCATTTG-1 1_779047_G_A 1 779047 0 G A 1 1 0|1
12 BS12_CATCCACCAGTGGTGA-1 1_779047_G_A 1 779047 0 G A 0 1 0|1
table(df_allele_filtered$CHROM)

  1      10      11      12      13      14      15      16      17      18      19       2      20      21      22       3       4       5       6 

1724619 908936 1024254 952319 496182 713317 638576 584814 884011 393062 701244 1449681 451849 269327 313893 1430917 1309189 1083317 1355959
7 8 9
1250710 1086883 1039017
Hi the pileup and phase run for 14 samples merged BAM was successful. I tried to set min_genes = 1, 5, 10, etc, the error came out like 10 genes for CHROM 18 or 5 genes for CHROM 21. Everything was OK with single sample run, I got all the final trees, subclones and graphs.