Low percentage of overlapping

[euduca]

Hello everyone.
I looked for some answers in the forum, but I still have doubts about overlapping.

I read about it on some sites and forums on the subject.
So I used the PEAR program to merge my reads.

This next generation sequencing procedure employed the Illumina platform HiSeq, the insert size was 800-900 bp and the library was TruSeq LT DNA kit. The average size of the sequences is from 100-101. (I used the FASTQC for the quality of information).

I first used the raw data and then the trimmed data (Trimmomatic).

In both cases, the overlapping percentage was very low.
So, this is good or bad?

Raw data:

PEAR v0.9.8 [April 9, 2015]

Citation - PEAR: a fast and accurate Illumina Paired-End reAd mergeR
Zhang et al (2014) Bioinformatics 30(5): 614-620 | doi:10.1093/bioinformatics/btt593

Forward reads file.................: ../../sequences/MP_CGATGT_L001_R1_001.fastq
Reverse reads file.................: ../../sequences/MP_CGATGT_L001_R2_001.fastq
PHRED..............................: 33
Using empirical frequencies........: YES
Statistical method.................: OES
Maximum assembly length............: 999999
Minimum assembly length............: 50
p-value............................: 0.010000
Quality score threshold (trimming).: 0
Minimum read size after trimming...: 1
Maximal ratio of uncalled bases....: 1.000000
Minimum overlap....................: 10
Scoring method.....................: Scaled score
Threads............................: 9

Allocating memory..................: 200,000,000 bytes
Computing empirical frequencies....: DONE
  A: 0.266503
  C: 0.233513
  G: 0.233699
  T: 0.266286
  2441195 uncalled bases
Assemblying reads: 100%

[B]Assembled reads ...................: 445,709 / 42,848,431 (1.040%)
Discarded reads ...................: 4,585 / 42,848,431 (0.011%)
Not assembled reads ...............: 42,398,137 / 42,848,431 (98.949%)[/B]
Assembled reads file...............: MP_CGATGT_L001.assembled.fastq
Discarded reads file...............: MP_CGATGT_L001.discarded.fastq
Unassembled forward reads file.....: MP_CGATGT_L001.unassembled.forward.fastq
Unassembled reverse reads file.....: MP_CGATGT_L001.unassembled.reverse.fastq

Trimmed data:

PEAR v0.9.8 [April 9, 2015]

Citation - PEAR: a fast and accurate Illumina Paired-End reAd mergeR
Zhang et al (2014) Bioinformatics 30(5): 614-620 | doi:10.1093/bioinformatics/btt593

Forward reads file.................: ../../../trim/trimmomatic/MP_CGATGT_L001_R1_001_p.fq
Reverse reads file.................: ../../../trim/trimmomatic/MP_CGATGT_L001_R2_001_p.fq
PHRED..............................: 33
Using empirical frequencies........: YES
Statistical method.................: OES
Maximum assembly length............: 999999
Minimum assembly length............: 50
p-value............................: 0.010000
Quality score threshold (trimming).: 0
Minimum read size after trimming...: 1
Maximal ratio of uncalled bases....: 1.000000
Minimum overlap....................: 10
Scoring method.....................: Scaled score
Threads............................: 9

Allocating memory..................: 200,000,000 bytes
Computing empirical frequencies....: DONE
  A: 0.267372
  C: 0.232951
  G: 0.231743
  T: 0.267933
  6664 uncalled bases
Assemblying reads: 100%
[B]
Assembled reads ...................: 380,009 / 32,256,001 (1.178%)
Discarded reads ...................: 0 / 32,256,001 (0.000%)
Not assembled reads ...............: 31,875,992 / 32,256,001 (98.822%)[/B]
Assembled reads file...............: MP_CGATGT_L001_trim.assembled.fastq
Discarded reads file...............: MP_CGATGT_L001_trim.discarded.fastq
Unassembled forward reads file.....: MP_CGATGT_L001_trim.unassembled.forward.fastq
Unassembled reverse reads file.....: MP_CGATGT_L001_trim.unassembled.reverse.fastq

Thank you, and I'm studying as much as I can about it.

P.S: Yes, I read the http://seqanswers.com/forums/showthread.php?t=66830 before.

[GenoMax]

How do you expect reads (~100 bp) sampled from two ends of a fragment (that is 800-900 bp) to overlap/merge?

Perhaps you should be aligning these reads to a reference rather than trying to overlap them directly?

BTW: Reads that are overlapping/merging likely represent cases where the insert (fragment) must be very short.