Table of Contents

• Introduction
• Tools
• Installation
• Usage
• Use u4falign for manipulating alignment results
• Difference between read-SAM and frag-SAM

Introduction

Falign is a sequence alignment toolkit for long noisy chromosome conformation capture (3C) reads, such as Pore-C. Falign is written in C and C++ programming language. Falign relies on the HTSLIB library. We have dowanloaded one copy of that and save in ./htslib/htslib-1.19.1.tar.bz2 to make it convenient for off-line installation.

Tools

Two tools are released together in this toolkit.

• falign. The alignment tool.
• dip3d. A utility used internally by Dip3D, a diploid human 3D genome construction algorithm.

Installation

Step 1: Download source codes

$ git clone https://github.com/xiaochuanle/Falign.git

Step 2: Install HTSLIB

$ cd Falign
$ ./htslib/install-htslib.sh

Step 3: Export HTSLIB to an environment variable

In Step 2 above, we install HTSLIB in directory ./htslib/htslib:

$ ls ./htslib/htslib
bin  include  lib  share

We have to tell Falign where to link HTSLIB. To do this, we export the location of this directory to an environment variable name LIBHTS:

$ export LIBHTS=$(pwd)/htslib/htslib
$ echo $LIBHTS
/share/home/chuanlex/chenying/data/Falign/htslib/htslib

Step 4: Install Falign

Before installing Falign, make sure you have executed Step 3 above. Otherwise the linker will complain that it cannot find -lhts.

$ cd src/ && make && cd ..

Falign is installed in the directory Linux-amd64/bin/:

$ ls Linux-amd64/bin/
dip3d  falign

Usage

Sample reads and reference

The directory ara-sample-data/ provides sample reference file reference.fa.gz and sample reads file reads.fq.gz.

Step 1: Build repeat regions for the reference genome

./Linux-amd64/bin/falign build-repeat ara-sample-data/reference.fa.gz ara-sample-data/reference.fa.gz.repeat.bed

Step 2: Map reads to the reference

./Linux-amd64/bin/falign -repeat_bed ara-sample-data/reference.fa.gz.repeat.bed -num_threads 4 ara-sample-data/reference.fa.gz ara-sample-data/reads.fq.gz > map.paf

The mapping results are output to the file map.paf in PAF format.

Output format

Falign supports different output formats (specified by the -outfmt option):

• PAF
• SAM
• BAM
• FRAG-SAM
• FRAG-BAM

In each output result (note that every alignment has for offsets: read start, read end, reference start, reference end), falign adds the following additional fields:

• qS:i: the nearest restiction enzyme site to the read start position
• qE:i: the nearest restiction enzyme site to the read end position
• vS:i: the nearest restriction enzyme site to the reference start position
• vE:i: the nearest restriction enzyme site to the reference end position
• pi:f: percentage of identity of the alignment
• SA:Z: a homologous map of the fragment

Difference between read-SAM and frag-SAM

By convention, in SAM mapping results, a read may contain multiple mapping results. For saving space, the read sequence is usually only presented in the first mapping result of this read. In the second to last mapping results of this read, the sequence field is filled by a start *. falign outputs SAM mapping results in this manner. And we call SAM mapping results output in this manner read-SAM. Since a Pore-C read always contains multiple fragments, a Pore-C read usually contains many mapping results:

0cd79600-51cf-4255-a6f7-0e9660721e85	16	2	1794202	1 ...
0cd79600-51cf-4255-a6f7-0e9660721e85	16	2	1791136	60 ...
0cd79600-51cf-4255-a6f7-0e9660721e85	0	2	1733063	60 ...
0cd79600-51cf-4255-a6f7-0e9660721e85	16	4	12727850	60 ...
0cd79600-51cf-4255-a6f7-0e9660721e85	0	2	1713207	60 ...

In the example above, the read 0cd79600-51cf-4255-a6f7-0e9660721e85 contains five alignment results.

Some tools such as whatshap taking BAM format as input will complain if there exist too many duplicate read names. In this case we suggest transfer read-SAM to frag-SAM. In frag-SAM every fragment is treadted as an individual read. We can use the sam2frag-sam command in the u4falign to transfer read-SAM to frag-SAM:

$ u4falign sam2frag-sam map.sam > frag-map.sam

After transformation, we have

0cd79600-51cf-4255-a6f7-0e9660721e85_0000000001:000:0000000000:0001794201	16	2	1794202	1 ...
0cd79600-51cf-4255-a6f7-0e9660721e85_0000000001:001:0000000000:0001791135	16	2	1791136	60 ...
0cd79600-51cf-4255-a6f7-0e9660721e85_0000000001:002:0000000000:0001733062	0	2	1733063	60 ...
0cd79600-51cf-4255-a6f7-0e9660721e85_0000000001:003:0000000001:0012727849	16	4	12727850	60 ...
0cd79600-51cf-4255-a6f7-0e9660721e85_0000000001:004:0000000000:0001713206	0	2	1713207	60 ...

In frag-SAM, the sequence field in every alignment results is represented by the fragment sequence (not the whose read sequence). Note that we add a suffix to every read name to avoid duplicate read names in the SAM file. The meanings of the fields in the suffix string is

read-id:fragment-id:reference-sequence-id:reference-mapping-position

Maintainers

• Chuan-Le Xiao, [email protected]
• Ying Chen, [email protected]

Citation

License

GPLv3