# Version 2
# Pandas/DataFrame/TSV-oriented, rather than FASTA-oriented. 
# Used in HTNA 2024

[DEFAULT]
dask_temp = ./temp

[project]
project_id=M300
project_leader=Matt Jacobs
project_institution=UCSC


[barcodes]
#ssifile = ~/git/mapseq-processing/etc/barcode_v2.txt
#ssifile = ~/git/mapseq-processing/etc/ssi_v2.txt
ssifile = ~/git/mapseq-processing/etc/novaseq_all.txt


[fastq]
r1start = 0
r1end = 32
# include 4nt nextseq rtag
#r1end = 36
# include 5nt novaseq rtag
#r1end = 37

r2start = 0
r2end = 20

# include 4nt nextseq rrtag
#r2end = 24

# include 5nt nextseq/novaseq rrtag
#r2end = 25

# to remove suspect sequences with long homopolymer runs. Should be < 1%?
max_repeats=7
# remove sequences with ambiguous bases
max_n_bases = 0
# reporting intervals for verbose output, defines how often to print. 
seqhandled_interval = 1000000

# Pandas number of lines to read from TSV at a time.
# Typically 1.5GB + 1.5GB (R1 + R2) for MAPseq FASTQs  
chunksize=50000000

# fast storage large enough for all data, e.g. ~500GB+
use_dask = False
dask_temp = ~/scratch

# unconditional threshold applied at aggregation step.
# shouldn't be needed. 
min_reads = 1

# applied to isolated filename for source column. 
# match everything up to first underscore. 
source_regex=(.+?)_


[split]
# arbitarary column creation from string.
# standard python slice semantics 
# Standard fields 52nt, no read tags
vbc_read_st = 0
vbc_read_end = 30
spikeseq_st=24
spikeseq_end = 32
libtag_st=30
libtag_end=32

# standard nextseq/novaseq fields
umi_st = 32
umi_end = 44
ssi_st = 44
ssi_end = 52

# include nextseq 4nt field GTAC, 4nt R2 CACG
#rtag_st=32
#rtag_end=36
#umi_st = 36
#umi_end = 48
#ssi_st = 48
#ssi_end = 56
#rrtag_st = 56
#rrtag_end = 60


# include novaseq 5nt field GTACT, 5nt R2 CACGA
#rtag_st=32
#rtag_end=37
#umi_st = 37
#umi_end = 49
#ssi_st = 49
#ssi_end = 57
#rrtag_st = 57
#rrtag_end = 62

[readfilter]
# filter read fields that don't match requirement
#
# novaseq 5nt
#rtag_seq = GTACT
#rrtag_seq = CACGA

# nextseq 4nt rtag
# rtag_seq = GTAC

# nextseq 4nt rrtag
# rrtag_seq = TGTA

drop_mismatch = False


[collapse]
tool = bowtie2
n_bases = 30
max_mismatch = 3
seq_length = 30
# Should track overall data size (unique VBC sequences). 
# 5000 for desktop/laptop, 40000+ for novaseq?
# max_recursion = 5000
max_recursion = 40000


[bowtie2]
threads = 10


[readtable]
# Readtable creation is where sample metadata is mapped/added:
#
# Sampleinfo.xlsx/tsv site-type definitions:
#
# LABEL						SOURCE				MEANING
# target					user-provided		treated sample
# target-negative           user-provided    	treated but expected to be low
# 	Both handled exactly the same and included in matrices. 

# injection					user-provided 	treated sample

# CONTROLS
# target-negative-control   user-provided   	untreated sample
# target-wt-control         core added      	untreated biological sample
# target-water-control      core added      	empty sample
# injection-water-control   core added      	empty sample

# 
# T or C = YY
# last 2nt of spike-ins AND reals
# A or G = RR
# last 2nt of L1 controls 
spikeseq= CGTCAGTC
realregex = [TC][TC]$
loneregex = [AG][AG]$

# Whether to identify L2 and L1 by libtag
use_libtag = True


[vbctable]
inj_min_reads = 2
target_min_reads = 2

# string evaluated, by eval(), to a list of integer tuples.
# (inj_min_umis, target_min_umis) 
# if populated, will create XLSX report of n_unique VBCs
# at the various thresholds.
# test_params = [ ]
test_params = [ (5,3) ,(10,3), (10,5), (20,5), (30,5),(30,10)]


[vbcfilter]
#  Require that target VBCs be present in the injection area to count.
#  Typically True, unless the dataset doesn't have any injection areas analyzed.
#  Must be false if there are no injection samples.
require_injection=False

# Include injection area(s) in matrix. Only VBCs also in targets, though.  
include_injection=False

# Include non-brain-specific controls in filtered table, so they will
# be included in matrices. 
include_controls = True

# Per-brain, threshold UMI count in target areas to only those 
# more than the largest target-negative total in that brain.  
use_target_negative=False
use_target_water_control=False

# minimum unique molecules (UMIs) in injection to be kept: 30
inj_min_umi = 10

# Alternatively explicit minimum unique molecules (UMIs) 
# in *any* target area to retain VBC in all areas (even if less than value).
target_min_umi = 2

# Minimum unique molecules (UMI) 
target_min_umi_absolute = 1



[matrices]
#clustermap_logscale=log10
clustermap_scale=log2

# Include non-brain-specific controls in in every brain matrix. 
include_controls = True