Skip to content

Latest commit

 

History

History
177 lines (147 loc) · 4.67 KB

LabJournal.md

File metadata and controls

177 lines (147 loc) · 4.67 KB
Raw

Lab Journal 📖

2022

05.09, Mon

  • reran the real motifs categorizing
  • found out how to solve root issue of conda container: copy an env as a user instead of using the base env.

04.24, Mon

  • amarel maintenance

04.20, Wed

  • amarel maintenance

04.19, Tue

  • amarel maintenance

04.18, Mon

  • read literatures
  • amarel maintenance

04.15, Fri

  • sick

04.14, Thu

  • sick

04.13, Wed
#AssessPertX

  • read literatures

04.13, Wed
#AssessPertX

  • read literatures #scRNA-SeqVariability
  • updat manuscript

04.12, Tue
#AssessPertX

  • read literatures

04.12, Tue
#AssessPertX

  • removed n_gained vs n_lost analyses

04.11, Mon
#AssessPertX

  • delta n_motifs
    • overall
      • passed all 4 filters: stat tests among 3 pert
      • didn't pass all 4 filters: stat tests among 3 pert
    • positive n_motifs
    • negative n_motifs
      #scRNA-SeqVariability
  • figuring out SCT vs counts, variance change, only check altered genes
    #coding
  • trying conda+jupyterlab singularity recipes
    • jlab works
    • conda pending trying

04.05, Tue
#AssessPertX

  • activator or repressor
    • passed all 4 filters
    • didn't pass all 4 filters
  • rotation summary #coding
  • trying conda+jupyterlab singularity recipes

04.04, Mon
#AssessPertX

  • activator or repressor
    • log2FC_72h
    • func_72h

04.03, Sun
#AssessPertX

  • updated KW tests into trimmed one way ANOVA

04.03, Sun

  • built singularity of more packages installed #scRNA-SeqVariability
  • updated manuscript

04.02, Sat

  • built singularity container of rserver, and rserver + Bioconductor packages

04.01, Fri #AssessPertX

  • updated analyses: check activator and repressor

03.31, Thu
#scRNA-SeqVariability

  • updated manuscript until the SCT part #AssessPertX
  • updated analyses: how many WT.overlapped are removed?

03.30, Wed
#scRNA-SeqVariability

  • updated manuscript until the SCT part
    #AssessPertX
  • updated the successful/unsuccessful props, perturbed/not perturbed props, and overlapped/not overlapped props
    • found a mistake: "Perturbed or not" should be the reversed of "contain_Identical.PertX.MotifIDs.in.sequence"
    • now is fixed
  • 18:30PM: found everything lost in amarel home directory

03.29, Tue
#scRNA-SeqVariability

  • finished different cell numbers/type -- all methods finished
    • methods: 3, 7, 9, 11, 16
  • MWW test finished

03.28, Mon
#AssessPertX

  • continue adding plots showing both motif numbers and statistics #scRNA-SeqVariability
  • finished different cell numbers/type
    • methods: 13

03.27
#scRNA-SeqVariability

  • finished different cell numbers/type
    • methods: 1, 12, 14, 15
      #AssessPertX
  • slides
    • add plots showing both motif numbers and statistics (haven't finished)
    • substitute Fisher's exact tests with Chi-square tests

03.26
#scRNA-SeqVariability

  • calculating affinities for scenarios of different cell numbers/type:
    • condition 1: 100 cells/type; condition2: 100-1000 cells/type
    • methods: 1, 3, 7, 9, 11, 12, 13, 14, 15, 16
  • manuscript: finished simulation part

03.22
#AssessPertX

03.18
#scRNA-SeqVariability

  • method1 needs high RAM, set as 400G, with 5 cores
  • tried installing glmGamPoi
    • Error: zlib.h not defined when installing XVector #bug
      • Solution: manually define path of zlib.h to $HOME/miniconda3
        1. in source code of XVector
        2. in ./src/io_utils.c, line 16: include{zlib.h}
        3. change zlib.h to absolute path of zlib.h
        4. re-tar source code and install package from source #debug
    • Error: "cannot find -lz" #bug

03.17
#scRNA-SeqVariability

  • "connection error" when using mclapply #debug #bug
    • solutions:
      1. using absolute paths
      2. setwd() before writing out files
      3. alwayse setwd() before sourcing scripts
  • revised Rscritps of methods1-14 to avoid duplicating calculations
    • check available files: list.files(..., recursive = T)
    • get not available fils: not.avail = 1:1000 %>% .[!. %in% avail.files]
    • lapply only on not.avail files
  • adequate job configs for methods.job (avoid long queues)
    • nCPU <= 10, best <=8
    • mem <=500G, best <= 200G
  • method13, 100cell/type, batch303: #bug
    • submit issue to github
    • answer from developer: install package glmGamPoi

03.16
#scRNA-SeqVariability

  • Revised .jobs of method
    • partition=mem --> less queueing
    • more CPUs
    • larger RAM
  • waiting for the methodN.job to finish