Bumblebee

Data

• ~ 500B genome

Original

• 7 pairs of data files (paired ends) : lanes 1..3,5..8 (lane 4 wasn't used)

 Lane   Insert            ReadLen    #Reads 
 1      3K(2..6,avg 4K)   124        34,944,099 
 2      8K(7..9,avg 8K)   124        32,540,640

 3      500(450..600)     124        34,745,750  # gDNA
 5      500                          34,601,239
 6      500                          34,553,857
 7      500                          34,682,612
 8      500                          12,975,839

• Adaptors

 >circularizarion
 CGTAATAACTTCGTATAGCATACATTATACGAAGTTATACGA
 >circularizarion.revcomp
 TCGTATAACTTCGTATAATGTATGCTATACGAAGTTATTACG 
 >5
 GATCGGAAGAGCGGTTCAGCAGGAATGCCGAGATCGGAAGAGCGGTTCAGCAGGAATGCCGAGACCG
 >3
 CGGCATTCCTGCTGAACCGAGATCGGAAGAGCGTCGTGTAGGGAAAGAGTGT
 

Tasks to figure out

1. Erroneous reads/bases, which we need to correct or discard
2. GC bias, so we can compute a-stats properly
3. Redundancy in the long paired ends, which are lane 1 and lane 2.

• Used the 454 protocol to circularize the DNA for sequencing with the Illumina instrument.
  • Some reads will begin in the circularization adaptor and thus will have only one usable read
  • Some reads have a few bases of DNA sequence and hit the circularization adaptor right away
  • Most reads will have at least 36bp from each end before hitting the adaptor.
  • Many reads will not have any adaptor to trim (>125bp of DNA sequence at both ends of the adaptor)

Trimming

• Keep only the first 100bp (last 24 bp are anyway low qual) otherwise gatekeeper "Seg fault"
• Adaptor trimming:
  • Split data set in 1M read subsets
  • Align all subsets to adaptors

 nucmer -l 8 -c 16 -b 8 -g 8 adaptors.seq subset.seq -p subset
 delta-filter -l 16 -q subset.delta > subset.filter-q.delta

• • Identify CLV

 ~/bin/delta2clr53.pl -5 5,3 -minLen 64 -maxLen 100 < subset.filter-q.delta > subset.clv

• Location:

 /fs/szattic-asmg4/Bees/Bombus_impatiens

Assembly

• Trimming

 No OBT
 adaptors in the seqs

• Kmers

 meryl -Dh -s 0-mercounts/asm-C-ms22-cm1 >! 22mers.hist
 Found 3136399464 mers.
 Found 379123530 distinct mers.
 Found 201257394 unique mers.
 Largest mercount is 12006651; 90 mers are too big for histogram.

 countKmers.pl
 most frequent 42mer : CGTAATAACTTCGTATAGCATACATTATACGAAGTTATACGA ~ 20%  of the seqs : circularization adapter  
 

• Unitigger : max utg len=852bp

• Consensus after unitigger : 3 out of 129 jobs failed

• Location

 /fs/szdevel/dpuiu/SourceForge/wgs-assembler.030210/Linux-amd64/bin/runCA