Scripts for assembly of PE Illumina from pooled BAC/FACs
To run the assembly use the assemble_facs.py script:
usage: assemble_facs.py [-h] -v VECTOR -1 FORWARD -2 REVERSE [-c CPUS]
Script to quality trim, subtract vector reads, and assemble using unicyler
optional arguments:
-h, --help show this help message and exit
-v VECTOR, --vector VECTOR FASTA vector sequence (default: None)
-1 FORWARD, --forward FORWARD FASTQ R1 (default: None)
-2 REVERSE, --reverse REVERSE FASTQ R2 (default: None)
-c CPUS, --cpus CPUS Number of CPUS (default: 8)
--skip_quality Skip quality/adapter trimming (default: False)
This would be run like so:
$ assemble_facs.py -v pFACvector.fasta -1 C1_S9_L001_R1_001.fastq.gz -2 C1_S9_L001_R2_001.fastq.gz
The output will be an assembly file called C1_unicycler.fasta.
After generating assembly for each pooled run, using a tab-delimited mapping file, can pull out full-length FACs.
The tab-delimited mapping file looks like this (The sanger reads are truncated in this example):
#Organism TeloFAC_Order FAC_Name Sanger_fwd_seq Sanger_rev_seq Column_Pool Row_Pool Plate_Pool Illumina_seqpools
Aspergillus aculeatus 1A1 IGAaBAC1D1 TAGGGCCGGTCGCGGGGGTCCCTG TTCAATAGCCCTATGGGTCGTGCG C1 RA P1 "C1, RA, P1"
This data can then be used to map the Sanger reads to the assembly using the get_full_length_facs.py script.
usage: get_full_length_facs.py [-h] -a ASSEMBLY -i INPUT [-p POOL [POOL ...]]
[--min_len_complete MIN_LEN_COMPLETE]
Script that parses tsv file, maps end sequences to FAC assembly.
optional arguments:
-h, --help show this help message and exit
-a ASSEMBLY, --assembly ASSEMBLY Assembly in FASTA format (default: None)
-i INPUT, --input INPUT TSV FAC end sequences (default: None)
-p POOL [POOL ...], --pool POOL [POOL ...] Name of seq pool (default: None)
--min_len MIN_LEN Minimum contig length to find end primers (default: 20000)
It can be run using the above example like this:
$ get_full_length_facs.py -a C1_unicycler.fasta -i IGteloFACpooldataNF.txt -p C1
This will generate 3 files:
- Raw Mapping data from minimap2
- FASTA file containing full length sequences
- FASTA file containing all sequences
The raw data looks like:
FAC_name read_num query_len query_start query_end query_cov strand contig contig_len ref_start ref_end match_len aln_len pident map_qual cigar
IGAwBAC1B10 1 1080 183 1078 82.9 -1 ctg00017 90870 89987 90870 875 895 97.76536312849163 60 9M1I20M1I5M1I5M1I12M1I22M1I8M1I19M1I65M1I7M1I71M1I23M1I617M
IGAwBAC1B9 1 985 60 983 93.7 1 ctg00007 130612 0 908 901 928 97.09051724137932 60 4M1I102M5D605M1I50M1I10M1I9M1I13M1I19M1I20M1I8M2I3M1I6M1I8M1I3M1I5M1I6M2I5M1I6M1I11M1I10M
IGAwBAC1B9 2 200 0 200 100.0 -1 ctg00007 130612 99649 99848 199 200 99.5 60 167M1I32M
IGAwBAC1C16 1 160 20 160 87.5 1 ctg00020 49887 23726 23866 140 140 100.0 60 140M
IGAwBAC1D11 1 200 15 200 92.5 1 ctg00020 49887 7514 7699 185 185 100.0 60 185M
IGAwBAC1D15 1 200 19 200 90.5 -1 ctg00008 129517 105945 106126 178 181 98.34254143646409 60 181M
IGAwBAC1D15 2 852 0 732 85.9 1 ctg00008 129517 36 755 705 732 96.31147540983606 60 585M1I21M1I34M1I9M1I5M1I17M1I9M1I11M1I5M1I7M2I10M2I6M
IGAwBAC1
And the "classified" FASTA file looks like this (if multiple FACS are overlapping separated by '|'):
>>IGAwBAC1B9;organism=Aspergillus wentii;origin=ctg00007;length=99848;full_length=yes;
AGCCCTAAGCCAAGCCCTAAGCCCTAAGCCCTAAGCCCTAAGCCCTAAGCCCTAAGCCCTAA...
>IGFaBAC4L14;organism=Fusarium solani;origin=ctg00038-ctg00011;length=146644;full_length=stitched;
GGATTCCGGAATGTGGAACGGGGCACAGGCACCGACCCTATCAAGGGGTAAGGGGATTCGGTCA...
>IGPmBAC1A7;organism=Penicillium marneffei;origin=ctg00013;length=109916;full_length=no;
AAACCCTAAACCCTAAACCCTAAACCCTAAACCCTAAACCCTAAACCCTAAACCCTAACCCTAAAC...
Since the pools are split over several samples, I suggest to run each pool separately first, then pull out full-length sequences for each one of those pools. You can then run a second or third round of assembly by subtracting the reads that map to the full-length sequences.
A simple example:
#assemble C1, RA, and P1 separately
$ assemble_facs.py -v pFACvector.fasta -1 C1_S9_L001_R1_001.fastq.gz -2 C1_S9_L001_R2_001.fastq.gz
$ assemble_facs.py -v pFACvector.fasta -1 RA_S1_L001_R1_001.fastq.gz -2 RA_S1_L001_R2_001.fastq.gz
$ assemble_facs.py -v pFACvector.fasta -1 P1_S21_L001_R1_001.fastq.gz -2 P1_S21_L001_R2_001.fastq.gz
#now pull out full length for each
$ get_full_length_facs.py -a C1_unicycler.fasta -i IGteloFACpooldataNF.txt -p C1
$ get_full_length_facs.py -a RA_unicycler.fasta -i IGteloFACpooldataNF.txt -p RA
$ get_full_length_facs.py -a P1_unicycler.fasta -i IGteloFACpooldataNF.txt -p P1
#combine all full-length sequences
$ cat *_fulllength-contigs.fasta > full-length_sequences.fasta
#combine the quality trimmed and vector subtracted data for each forward/reverse read
$ cat C1_clean_R1.fastq.gz RA_clean_R1.fastq.gz P1_clean_R1.fastq.gz > cleaned_R1.fastq.gz
$ cat C1_clean_R2.fastq.gz RA_clean_R2.fastq.gz P1_clean_R2.fastq.gz > cleaned_R2.fastq.gz
#we can now re-run assembly but subtract reads that are already part of full-length sequences
$ assemble_facs.py -v full-length_sequences.fasta -1 cleaned_R1.fastq.gz -2 cleaned_R2.fastq.gz
#and then finally pull out full-length from this combined assembly
$ get_full_length_facs.py -a q-trimmed_unicycler.fasta -i IGteloFACpooldataNF.txt --pool C1 RA P1
There will still be some ~100 kb inserts that don't seem to match Sanger end sequences - they can be searched against the reference genome to figure out where they came from. Then there is also an accessory script that you can use to filter the FASTA files, which can be run like this:
#to ouput all sequences > 80kb
filter_seqs.py -i P4.all-contigs.fasta -l 80000
#to ouput all sequences > 80kb that are Aspergillus wentii
filter_seqs.py -i P4.all-contigs.fasta -m 80000 -s "Aspergillus wentii"
#to ouput the longest unique contigs
filter_seqs.py -i P4.all-contigs.fasta -l
#to ouput the longest unique contigs from Fusarium solani
filter_seqs.py -i P4.all-contigs.fasta -l -s "Fusarium solani"
- Python 2
- biopython
- trim_galore
- minimap2
- unicycler
- samtools
- mappy (minimap2 python library
React
Typescript
Django
Laravel
Facebook
Microsoft
Google
Alibaba
D3
Tencent