config.yaml

# 0 general info ----------------------------------------------------------------------------------

title: Title of my project

authors:
  - first_author
  - second_author
  - pi_1
  - pi_2

# number of cores to utilize per node.
# never set more cores than allowed per node (usually 24 but check the hardware specs).
cores: 8

# tracks the beginning of an analysis.
# This value is preserved across all stages of the analysis.
date: "2021-02-22" # YYYY-MM-DD

# list out antibodies exactly as how they are labeled from cell ranger
# if antibodies are labeled with underscores, convert them to dashes like below.
antibody_list:
  - ab-CD10
  - ab-CD34
  - ab-CD38
  - ab-CD45RA
  - ab-CD90
  - ab-CD99
  - ab-CD123
  - ab-CD366

assembly: "hg38"

# 1 preprocessing ---------------------------------------------------------------------------------

# list the names of your samples to process.
samples2process:
  - dmso
  - ory

# give the analysis a name.
project_name: "my_project"

# save or don't save seurat object with cell cycle regression
cell_cycle_regression: TRUE # [TRUE, FALSE]

# See QC metrics at these percentiles.
metadataQuants:
  - 0
  - 0.25
  - 0.5
  - 0.75
  - 0.95
  - 1

# filtering criteria
percentMitoFilt: 15
nCountMinFilt: 2000
nCountMaxFilt: 20000

# principle components to calculate for each sample.
nPCs: 150

# principle components to utilize for each sample.
usePCs: 20

# visualize cell cycle states versus expression level of these cell cycle genes.
cc_genes:
  - NASP
  - USP1
  - TUBB4B
  - HMGB2
  - PCNA
  - TOP2A
  - MCM6
  - MKI67

# 2 integration -----------------------------------------------------------------------------------

# Set the baseline condition
baseline: "dmso"

# number of principal components to use for integrating transcriptome features. Should be <= usePCs
integration_anchor_PC: 20

# by default, the pipeline will integrate all PCs in the ADT space!

# how many neighbors (k) to use when picking anchors
k.anchor: 5

# 3 clustering- -----------------------------------------------------------------------------------

# the resolution value to cluster on per sample and save to the Seurat object.
# value should be a float between 0.1 and 2 (inclusive) with step size of 0.1
resolutions:
  dmso: 0.3
  ory: 0.3
  integrated: 0.2

# List of genes to view gene expression in UMAP space
FeaturePlotList:
  - PROCR
  # - HOXA9
  - CDK6
  - SOX4
  - MPO
  - ELANE
  - CAMP
  - CEBPE
  - LYZ
  - IRF8
  # - IL2RB
  - JCHAIN
  - HBB-BT
  - CAR1
  - GATA2
  - VCAN
# letter casing counts! if you don't see your gene in the output, try a different casing.
# default gene list reference: https://www.biorxiv.org/content/10.1101/2020.09.14.295295v1

# select a rcolorbrewer palette for cluster colors.
# Ideally use a qualitative or diverging color scheme.
rcolorbrewer_palette: "Spectral"

# 4 Differential Testing + Ontology ---------------------------------------------------------------

# Find all DE genes for each cluster. This will take a long time.
FindAllMarkers: FALSE # [TRUE, FALSE]

# Find markers conserved in each cluster, split by a factor (e.g. identity).
FindConservedMarkers: FALSE # [TRUE, FALSE]

# Find DE genes between two identities for each cluster in the integrated object. This is quick.
FindMarkersIdent: TRUE # [TRUE, FALSE]

# all the above DE schemes follow the following configuration
FindMarkersConfig:
  group1: "ory"
  group2: "dmso"
  grouping.var: "orig.ident"

# significance level for DE and GO
significance: 0.05

# filter genes using regex or FALSE for gene ontology. Case will be ignored. example:
# "^RPS|^RPL" will remove ribosomal S and L genes
filter_genes: FALSE

# common mistakes
# mismatch of gene casing.
# too few cells so DE breaks
# if you are experiencing DE problems, you can quickly debug by down-sampling cells in chunk3.