Only about 20% reads are mapped using bfast

[Jerry-cs]

Hi all,

I used bfast+bwa. For the F3 reads, only about 20% reads are mapped using bfast match. Is this number good enough?

Below is the output from bfast match
************************************************************
Printing Program Parameters:
programMode: [ExecuteProgram]
fastaFileName: /share/data9/genomes/human_all.fasta
mainIndexes [Auto-recognizing]
secondaryIndexes [Not Using]
readsFileName: Pla0000325656_1_PE_HS26621_1_F3_.15.fastq
offsets: [Using All]
loadAllIndexes: [Not Using]
compression: [Not Using]
space: [Color Space]
startReadNum: 1
endReadNum: 2147483647
keySize: [Not Using]
maxKeyMatches: 8
maxNumMatches: 384
whichStrand: [Both Strands]
numThreads: 8
queueLength: 250000
tmpDir: /share/data11/solid/tmp/
timing: [Using]
************************************************************
Searching for main indexes...
Found 4 index (4 total files).
Not using secondary indexes.
************************************************************
Reading in reference genome from /share/data9/genomes/human_all.fasta.cs.brg.
In total read 25 contigs for a total of 3080436051 bases
************************************************************
Reading Pla0000325656_1_PE_HS26621_1_F3_.15.fastq into a temp file.
Will process 1000000 reads.
************************************************************
Searching index file 1/4 (index #1, bin #1)...
Reading index from /share/data9/genomes/human_all.fasta.cs.1.1.bif.
Read index from /share/data9/genomes/human_all.fasta.cs.1.1.bif.
Reads processed: 1000000
Cleaning up index.
Searching index file 1/4 (index #1, bin #1) complete...
Found 185475 matches.
************************************************************
Searching index file 2/4 (index #2, bin #1)...
Reading index from /share/data9/genomes/human_all.fasta.cs.2.1.bif.
Read index from /share/data9/genomes/human_all.fasta.cs.2.1.bif.
Reads processed: 1000000
Cleaning up index.
Searching index file 2/4 (index #2, bin #1) complete...
Found 230886 matches.

...

[westerman]

Good enough for what? We need to know your experimental design. E.g., transcriptome project? whole genome resequencing? environmental sampling?

Depending on the project I've had people who were ecstatic that we obtained 3% mapping and people who were not happy with 75% mapping.

If I was to make a wild guess, since you are working with human and SOLiD data, I would say that 20% is too low. But we do need to know what you are trying to accomplish before anything definitive can be said.

[Jerry-cs]

Thanks, the reads are from human genome sequencing data. I just checked the data and found the poor mapping rate was caused by the high error rate of the data. When we use better quality data, the mapping rate is more than 65%.

Good enough for what? We need to know your experimental design. E.g., transcriptome project? whole genome resequencing? environmental sampling?

Depending on the project I've had people who were ecstatic that we obtained 3% mapping and people who were not happy with 75% mapping.

If I was to make a wild guess, since you are working with human and SOLiD data, I would say that 20% is too low. But we do need to know what you are trying to accomplish before anything definitive can be said.

[Chipper]

Why are you only using 4 indexes instead of the recommended 10? If your bad colors are periodic (eg every 5th color is missing) you may want to design a new index that compensates for it to get decent mapping rates.

[JohnK]

Good enough for what? We need to know your experimental design. E.g., transcriptome project? whole genome resequencing? environmental sampling?

Depending on the project I've had people who were ecstatic that we obtained 3% mapping and people who were not happy with 75% mapping.

If I was to make a wild guess, since you are working with human and SOLiD data, I would say that 20% is too low. But we do need to know what you are trying to accomplish before anything definitive can be said.

You couldn't have said it better. There is a huge disparity between Biologists' 'ideals' and the realities of mapping throughput.