FALCON-integrate - millanek (Sanger) fork

This is a slightly modified version of the PacBio repository to allow more customisation, specifically to make it easier to run the FALCON assembler on the Wellcome Trust Sanger Institute's compute cluster.

set-up

git clone https://github.com/millanek/FALCON-integrate.git
cd FALCON-integrate
git submodule update --init

I suggest you install the Python code under your own Python installation: e.g.

cd FALCON
~/programs/Python-2.7.12/python setup.py install --user

To run the FALCON code you will also need to compile the modified DALIGNER code and have the folder in your PATH:

cd ../DAZZ_DB
make
cd ../DALIGNER
make
PATH=/path/to/FALCON-integrate/DALIGNER:$PATH

The FALCON_unzip pipeline here is in its original PacBio version - needs some modification to get it run at Sanger I have the pipeline running but it is not public on gitHub; please contact me directly (see below)

contact

This fork is maintained by Milan Malinsky Any issues with getting FALCON work at Sanger, please let me know at [email protected]

Some more information is on the PacBio wiki.

submodules

In case you are unfamiliar with git-submodules, they are quite easy to use from the command-line:

git submodule update --init

If that fails, you can update your git, or try this:

git submodule init
git submodule update

which is almost the same thing.

example FALCON assembly workflow (after complete DAZZLER pipeline):

these are just examples of parameters that worked reasonably well for one particular assembly; much more info on parameter tuning is here: https://github.com/PacificBiosciences/FALCON/wiki/Manual

error-correct

for i in {1..N}; do bsub -G yourGroup -o correct_o_%J -e correct_e_%J -R'select[mem>15000] rusage[mem=15000]' -M15000 sh -c 'LA4Falcon -H 5000 -fo DATABASE_patched.db DATABASE_patched.${1}.las | fc_consensus.py --output_multi --n_core 0 --min_cov 6 --max_cov_aln 60 --max_n_read 200 > DATABASE_patched.${1}.corrected_max_cov_aln60_multi.fasta' ec $i; done
for i in {1..N}; do echo $i; awk 'BEGIN{ FS ="_"; printThis = 0;}{ if (substr($1,1,1) == ">") { if ($2 > 7000) { printThis = 1; print;} else {printThis = 0;}} else { if (printThis == 1) {print;}} }' DATABASE_patched.${i}.corrected_max_cov_aln60_multi.fasta > DATABASE_patched.${i}.corrected_max_cov_aln60_multi_min7kb.fasta
/path/to/FALCON-integrate/renumber_fasta.sh DATABASE_patched.${i}.corrected_max_cov_aln60_multi_min7kb.fasta

set up new DAZZLER database of corrected reads and build multiple alignments

geting new .las files

overlap filtering:

for i in {1..N}; do echo DATABASE_patched.corrected_max_cov_aln60_multi_min7kb.${i}.las >> fofn.txt; done
bsub -G yourGroup -o overlapFilter_o_%J -e overlapFilter_e_%J -R'select[mem>25000] rusage[mem=25000]' -M25000 sh -c 'fc_ovlp_filter.py --db DATABASE_patched.corrected_max_cov_aln60_multi_min7kb.db --fofn fofn.txt --n_core 0 --min_cov 10 --max_cov 120 --bestn 10 --max_diff 90 > filtered_overlaps_DATABASE_patched.corrected_max_cov_aln60_multi_min7kb.ovlp'

do the assembly with different min length cutoffs:

for minl in 6000 7000 8000 9000 10000 11000 12000 13000 14000 15000 16000 17000; do bsub -G yourGroup -o graphBuild_o_%J -e graphBuild_e_%J -R'select[mem>16000] rusage[mem=16000]' -M16000 sh -c 'fc_ovlp_to_graph.py --min_len $1 --params_fn minl_${1} filtered_overlaps_DATABASE_patched.corrected_max_cov_aln60_multi_min7kb.ovlp; fc_graph_to_contig.py --run_name minl_${1} DATABASE_patched.corrected_max_cov_aln60_multi_min7kb_renumbered_onlyReadID.fasta' assemle $minl; done