MultiGPS is freezing

[Lucas446]

Hi,

I am trying to use MultiGPS to analyze ChIPseq data, 3 replicates and 2 conditions. It loads the pair-end bam files well, but keep freezing for ever at Round 0.

Input:
-Pair-end bam file (bowtie2 aligned on Drosophila dm6 reference)

Software:
-multigps_v0.75.jar
-R version 3.6.1 (2019-07-05)
-Java version "1.8.0_192"
-MacOS v10.13.6 (High Sierra)

Hardware:
-Intel Core i5, 3.4 GHz, 4 cores
-Memory: 32 GB

Command line:
java -Xmx20G -jar /Applications/bio_info_tools/multigps_v0.75.jar -out MultiGPS_DamID_ --geninfo /Users/tlucas/Desktop/MULTIGPS/dm6.chrom_Edited.size.genome.sizes --design /Users/tlucas/Desktop/MULTIGPS/Design.txt --nomotifs
MultiGPS version 0.75

_No genome sequence data was provided with --seq, so motif-finding and the motif prior are switched off.
Processing HitLoaders for: EARLY 1
Processing HitLoaders for: EARLY 2
Processing HitLoaders for: EARLY 3
Processing HitLoaders for: EARLY 1
Processing HitLoaders for: EARLY 2
Processing HitLoaders for: EARLY 3
Processing HitLoaders for: LATE 1
Processing HitLoaders for: LATE 2
Processing HitLoaders for: LATE 3
Processing HitLoaders for: LATE 1
Processing HitLoaders for: LATE 2
Processing HitLoaders for: LATE 3
Loading data from EARLY:1:control Loaded.
Loading data from EARLY:2:control Loaded.
Loading data from EARLY:3:control Loaded.
Loading data from EARLY:1:signal Loaded.
Loading data from EARLY:2:signal Loaded.
Loading data from EARLY:3:signal Loaded.
Loading data from LATE:1:control Loaded.
Loading data from LATE:2:control Loaded.
Loading data from LATE:3:control Loaded.
Loading data from LATE:1:signal Loaded.
Loading data from LATE:2:signal Loaded.
Loading data from LATE:3:signal Loaded.
Calculating scaling factors for condition: EARLY Complete.
Calculating scaling factors for condition: LATE Complete.
Loaded all experiments:
Condition EARLY: #Replicates: 3
Replicate: EARLY:1
Signal: 11754764.0 Control: 22351332.0 ScalingFactor: 0.644
Replicate: EARLY:2
Signal: 12428313.0 Control: 32652716.0 ScalingFactor: 0.470
Replicate: EARLY:3
Signal: 15635568.0 Control: 16897252.0 ScalingFactor: 0.943
Pooled replicates for condition: EARLY
Signal: 39818645.0 Control:71901300.0 ScalingFactor:0.648
Condition LATE: #Replicates: 3
Replicate: LATE:1
Signal: 6541607.0 Control: 15878219.0 ScalingFactor: 0.594
Replicate: LATE:2
Signal: 2377084.0 Control: 25501429.0 ScalingFactor: 0.121
Replicate: LATE:3
Signal: 25964005.0 Control: 15175740.0 ScalingFactor: 1.486
Pooled replicates for condition: LATE
Signal: 34882696.0 Control:56555388.0 ScalingFactor:0.674
Finding potential binding regions.
PotentialRegionFilter: genomic threshold for EARLY with bin width 368.0 = 185.0
PotentialRegionFilter: genomic threshold for LATE with bin width 368.0 = 165.0
22453 potential regions found.
Initialzing mixture model
Alpha EARLY Range=368 70.0
Alpha LATE Range=368 58.0
Global noise per base initialization for EARLY = 0.1067
Global noise per base initialization for LATE = 0.0823
Multi-condition positional priors: A given B:-0.1520 A given notB:-15.9658

============================ Round 0 ============================_

Thanks a lot for your help,
Best,
Tanguy

[shaunmahony]

Hi Tanguy. Have you let it continue to run? Unfortunately, MultiGPS' mixture modeling approach can take a long time to complete analysis, especially on datasets where there are expected to be many binding events. I would recommend running it on a cluster and give it access to ~8 processors. It should run overnight then. But let me know if you continue having issues.