In [1]:
# Parameters
sample_name = "CLP;A;CL"
modisco_root = "/srv/scratch/msharmin/mouse_hem/with_tfd/full_mouse50/Naive_modisco2019"
mitra_subdir = "report/version2"
task_dir = "task_94-naivegw"
database_name = "CISBP"
perf_file = "/srv/scratch/msharmin/mouse_hem/with_tfd/full_mouse50/fineFactorized/task_94-naivegw/NaiveauPRC.txt"
homer_root = "/srv/scratch/msharmin/mouse_hem/with_tfd/full_mouse50/Naive_scans"
In [2]:
from matlas.modisco_report import modisco_report_pipeline, display_metadata
reportfile= "/mnt/lab_data/kundaje/msharmin/mouse_hem/with_tfd/full_mouse50/filtering samples_MS2.xlsx"
sheetname = "filter23"

load data from labcluster

Using TensorFlow backend.
2019-07-22 19:56:40,923 [WARNING] git-lfs not installed
In [3]:
display_metadata(sample_name, perf_file, reportfile, sheetname)
    Sample Information
    MetaData NameDescription
    Cell typeCLP
    Cell GroupOther Immune Cells
    Experiment NameATAC
    Experiment GroupGEO
    Pipeline Output
    replicateNaïve overlap peaksIDR peaksTSS enrichment (< 8 is very poor <10 is low)Final number of unique mapping, dup-filtered, chrM filtered readsNumber of reads in called peak regionsFraction of reads in called peak regionsNumber of reads in promoter regionsFraction of reads in promoter regionsNumber of reads in enhancer regionsFraction of reads in enhancer regions
    rep11389128969912.2998119702408519980.07139653140.080839860630.3335
    rep21389128969911.82141682005932494400.193424282970.144568566710.4081
    rep31389128969912.25711552209829274010.188822907150.147763298060.4083
    Modelling Metadata
    MetricValue
    auPRC0.5653
    Calibrated Recall at 50% FDR0.258
    Number of Positive Examples in Test Data101808
    Number of Negative Examples in Test Data7969043
    Imbalance Ratio in Test Data0.0126
    Test Chromosomeschr2, chr3, chr19
In [4]:
from matlas.modisco_report import display_comparative_motif_sets
display_comparative_motif_sets(sample_name, homer_root, modisco_root)

TF-MoDISco is using the TensorFlow backend.
Number of CISBP motifs obtained by TF-MoDISco and Homer-denovo
Shared Motifs
Motif NameModiscoHomer
Sfpi1
Sp3
Ctcf
Bcl11a
Pbx3
Runx1
Unique TF-MoDISco Motifs
Motif NameModiscoHomer
Nfiaabsent
Restabsent
Hoxc9absent
Bach2absent
Creb3absent
Erfabsent
Smarcc2absent
Klf16absent
Mef2dabsent
Ergabsent
Mbtps2absent
Rfx1absent
Nfe2l2absent
Unique Homer Motifs
Motif NameModiscoHomer
Yy1absent
E4f1absent
Zfp143absent
In [5]:
modisco_report_pipeline(sample_name, modisco_root, mitra_subdir, task_dir, database_name, 
                        importance=True, render=True)

rsync -t -av /srv/scratch/msharmin/mouse_hem/with_tfd/full_mouse50/Naive_modisco2019/task_94-naivegw/cisbp_tomtomout /srv/www/kundaje/msharmin/report/version2/task_94-naivegw/
chmod -R +755 /srv/www/kundaje/msharmin/report/version2/task_94-naivegw
Displaying motifs which has positive importances for the cell type
metacluster_1, # patterns: 23, # seqlets: 26962, Positive for: CLP;A;CL
In [6]:
modisco_report_pipeline(sample_name, modisco_root, mitra_subdir, task_dir, database_name, 
                        importance=False, render=True)

rsync -t -av /srv/scratch/msharmin/mouse_hem/with_tfd/full_mouse50/Naive_modisco2019/task_94-naivegw/cisbp_tomtomout /srv/www/kundaje/msharmin/report/version2/task_94-naivegw/
chmod -R +755 /srv/www/kundaje/msharmin/report/version2/task_94-naivegw
Displaying motifs which has negative importances for the cell type
metacluster_0, # patterns: 18, # seqlets: 9445, Negative for: CLP;A;CL
  • pattern_0: # seqlets: 2215 Cisbp matches (q-value>=0.01) using Tomtom (Full Report) :

    Dbx1, Tlx2, Arid3b, Pou4f3, Pou4f1

    SequenceContrib ScoresHyp_Contrib Scores
  • pattern_1: # seqlets: 1671 Cisbp matches (q-value>=0.01) using Tomtom (Full Report) :

    Zbtb7a, Glis3, Rreb1, Klf16, Zfp281

    SequenceContrib ScoresHyp_Contrib Scores
  • pattern_2: # seqlets: 1485 Cisbp matches (q-value>=0.01) using Tomtom (Full Report) :

    Zeb1

    SequenceContrib ScoresHyp_Contrib Scores
  • pattern_3: # seqlets: 995 Cisbp matches (q-value>=0.01) using Tomtom (Full Report) :

    SequenceContrib ScoresHyp_Contrib Scores
  • pattern_4: # seqlets: 604 Cisbp matches (q-value>=0.01) using Tomtom (Full Report) :

    Ppara, Rarb, Rxra, Nr2f6, Rara

    SequenceContrib ScoresHyp_Contrib Scores
  • pattern_5: # seqlets: 561 Cisbp matches (q-value>=0.01) using Tomtom (Full Report) :

    Tcf15, Tal1, Atoh1, Bhlha15, Tcfap4

    SequenceContrib ScoresHyp_Contrib Scores
  • pattern_6: # seqlets: 296 Cisbp matches (q-value>=0.01) using Tomtom (Full Report) :

    Hsf1

    SequenceContrib ScoresHyp_Contrib Scores
  • pattern_7: # seqlets: 295 Cisbp matches (q-value>=0.01) using Tomtom (Full Report) :

    Nr5a2

    SequenceContrib ScoresHyp_Contrib Scores
  • pattern_8: # seqlets: 276 Cisbp matches (q-value>=0.01) using Tomtom (Full Report) :

    Runx1, Runx2

    SequenceContrib ScoresHyp_Contrib Scores
  • pattern_9: # seqlets: 239 Cisbp matches (q-value>=0.01) using Tomtom (Full Report) :

    Foxd3, Foxj1, Foxp4, Foxm1, Onecut2

    SequenceContrib ScoresHyp_Contrib Scores
  • pattern_10: # seqlets: 149 Cisbp matches (q-value>=0.01) using Tomtom (Full Report) :

    SequenceContrib ScoresHyp_Contrib Scores
  • pattern_11: # seqlets: 146 Cisbp matches (q-value>=0.01) using Tomtom (Full Report) :

    Hoxb1

    SequenceContrib ScoresHyp_Contrib Scores
  • pattern_12: # seqlets: 143 Cisbp matches (q-value>=0.01) using Tomtom (Full Report) :

    SequenceContrib ScoresHyp_Contrib Scores
  • pattern_13: # seqlets: 111 Cisbp matches (q-value>=0.01) using Tomtom (Full Report) :

    SequenceContrib ScoresHyp_Contrib Scores
  • pattern_14: # seqlets: 89 Cisbp matches (q-value>=0.01) using Tomtom (Full Report) :

    SequenceContrib ScoresHyp_Contrib Scores
  • pattern_15: # seqlets: 74 Cisbp matches (q-value>=0.01) using Tomtom (Full Report) :

    Bcl6

    SequenceContrib ScoresHyp_Contrib Scores
  • pattern_16: # seqlets: 53 Cisbp matches (q-value>=0.01) using Tomtom (Full Report) :

    Bcl6, Sfpi1

    SequenceContrib ScoresHyp_Contrib Scores
  • pattern_17: # seqlets: 43 Cisbp matches (q-value>=0.01) using Tomtom (Full Report) :

    Bcl6, Sfpi1, Bcl11a, Irf1, Spib

    SequenceContrib ScoresHyp_Contrib Scores