Dpuiu Assemblathon: Difference between revisions

From Cbcb
Jump to navigation Jump to search
Line 101: Line 101:
   [http://www.ncbi.nlm.nih.gov/sra/SRX033397?report=full SRX033397] pair lib ;    readLen=101 ; insMea=180  
   [http://www.ncbi.nlm.nih.gov/sra/SRX033397?report=full SRX033397] pair lib ;    readLen=101 ; insMea=180  
   [http://www.ncbi.nlm.nih.gov/sra/SRX016063?report=full SRX016063] jumping lib ; readLen=101 ; insMea~=3455; ~15% of the mates are short inserts (~250bp)
   [http://www.ncbi.nlm.nih.gov/sra/SRX016063?report=full SRX016063] jumping lib ; readLen=101 ; insMea~=3455; ~15% of the mates are short inserts (~250bp)
=== Assembly ===
[[Media:Rhodobacter_sphaeroides.assembly.summary]]


== Human, a single chromosome, medium-sized ==
== Human, a single chromosome, medium-sized ==

Revision as of 18:10, 15 March 2011

Links

GAGE

  • Location
 http://gage.cbcb.umd.edu/ -> /fs/web-cbcb-new/html/gage

Assemblers

* Allpaths-LG    /fs/szdevel/core-cbcb-software/Linux-x86_64/packages/allpaths3-35218/
* CA             /fs/szdevel/core-cbcb-software/Linux-x86_64/packages/wgs-6.1/
* Velvet         /fs/szdevel/core-cbcb-software/Linux-x86_64/packages/velvet_1.0.13/
* SOAPdenovo     /fs/szdevel/core-cbcb-software/Linux-x86_64/packages/SOAPdenovo_Release1.04

CBCB genomes

  • a bacterial genome. Instead of E. coli, we can use S. aureus USA300, which has sequence data in SRA from 454 and Illumina, paired and unpaired. Daniela has already assemblied it using CA, Newbler, Velvet, SOAPdenovo, and Maq (using its comparative assembly mode, where it aligns to a reference).
  • A medium-sized eukaryote. I'd like to use the Argentine ant or the Bombus impatiens bee - I've just written to Gene Robinson to ask about the bee.
  • Another eukaryote, ideally a larger one. Human would be great, but we just don't have enough time to do multiple human assemblies. So maybe another insect, or perhaps a plant if we can find one for which data is available.

If we can agree on the data sets, then the next step would be to design the experiment - decide in advance which assemblers to run and how many ways to try each one. I'm thinking we should also trim all the data with Quake.

Argentine ant

Bombus impatiens

Data

  • 497,318,144 Illumina 124bp reads
  • 8 libraries; inserts:
    • 400bp
    • 3k (outie)
    • 8k (outie)
  • Traces

Adapters: in 3k & 8k libraries

C CGTAATAACTTCGTATAGCATACATTATACGAAGTTATACGA
3 CGGCATTCCTGCTGAACCGAGATCGGAAGAGCGTCGTGTAGGGAAAGAGTGT
5 GATCGGAAGAGCGGTTCAGCAGGAATGCCGAGATCGGAAGAGCGGTTCAGCAGGAATGCCGAGACCG
  • Locations:
/fs/szattic-asmg4/Bees/Bombus_impatiens/s_[12356789]_[12]_sequence.txt                             # original fastq files

/fs/szattic-asmg4/Bees/Bombus_impatiens/error_free/fastq/s_[129]_[012]_sequence.cor.rev.txt        # adaptor free corrected reads (long inserts)
/fs/szattic-asmg4/Bees/Bombus_impatiens/error_free/fastq/s_[35678]_[012]_sequence.cor.txt          # corrected reads (short inserts)

Assembly

/fs/szattic-asmg5/Bees/Bombus_impatiens/Assembly/CA.s_1-8.cor.redo2/                               # best Celera Assembly
/fs/szattic-asmg5/Bees/Bombus_impatiens/Assembly/SOAPdenovo.s_1-9.cor/                             # best SOAPdenovo assembly

Staph aureus USA300

Data

  • Complete genome  : NC_010079 2872915bp Staphylococcus aureus subsp. aureus USA300_TCH1516
 SRP001086 Staphylococcus aureus Sequencing on Illumina
 SRX007714 pair lib
 SRX007711 jumping lib
  • Locations:
 /nfshomes/dpuiu/HTS/Staphylococcus_aureus/Data/Illuminap100/
 /nfshomes/dpuiu/HTS/Staphylococcus_aureus/Data/Illuminaj/

Assembly

 ~dpuiu/GAGE/Staphylococcus_aureus/
 /fs/szattic-asmg5/dpuiu/HTS/Staphylococcus_aureus/Illumina.180_45X
 /fs/szattic-asmg5/dpuiu/HTS/Staphylococcus_aureus/Illumina.180_45X.3500_45X/
  • SOAPdenovo v1.05 :
    • new quake version did not help much (quake-0.2.2 vs davek44-error_correction-28dbe11)
    • SOAPdenovo map -K 37+ : fails on quakeCor.k18 corrected reads
    • "according" to kmerFreq , should probably not use -K >47
    • longer kmer => longer scaffolds (K=63 : largest N50scf)
    • longer kmer => shorted contigs (K=31 : largest N50ctg)
    • K40+ too large: no "valley" in the kmerFreq histogram
 paste SOAPdenovo.K??.quakeCor.k18/genome.K??.kmerFreq | nl0 | head
 paste SOAPdenovo.K??.allpathsCor/genome.K??.kmerFreq | nl0 | more

Rhodobacter sphaeroides

Data

  • Complete genome: Rhodobacter sphaeroides 2.4.1 : 2 chromosomes, 5 plasmids
 CP000143    3188609  
 CP000144    943016   
 DQ232586    114045   
 CP000145    114178   
 CP000146    105284   
 CP000147    100828   
 DQ232587    37100    
 Total       4603060
 SRX033397 pair lib ;    readLen=101 ; insMea=180 
 SRX016063 jumping lib ; readLen=101 ; insMea~=3455; ~15% of the mates are short inserts (~250bp)

Assembly

Media:Rhodobacter_sphaeroides.assembly.summary

Human, a single chromosome, medium-sized

Data

  • Latest online assembly
 ftp://ftp.ncbi.nih.gov/genomes/H_sapiens/Assembled_chromosomes/seq/
  NC_000014.8    107,349,540  # total, with telomeric N's 
                  88,289,540  # clean
  • Human bowtie indexes
  /fs/szdata/bowtie_indexes/h_sapiens_37_asm
  • Illumina reads (all genome)
 Human NA12878 Genome on Illumina 
 ftp://ftp-trace.ncbi.nlm.nih.gov/sra/sra-instant/reads/ByStudy/litesra/SRP/SRP003/SRP003680/
 ginko:/scratch1/Human_NA12878_on_Illumina/
 #Fragment (mean insert size: 155bp, SD 26), 101 bp read length
 Lib          #Spots  #Bases  #Reads     #Mates     ReadLen  InsMea  InStd  InsMin  InsMax   TrimReadLen
 SRR067787    82.4M   16.6G   652448124  324283604  101      155     26     77      458     
 SRR067789    82.6M   16.7G   654133372  324876520  101      155     26     77      458     
 SRR067780    83.3M   16.8G   660001672  328021140  101      155     26     77      458     
 SRR067791    83.0M   16.8G   657963460  327205952  101      155     26     77      458     
 SRR067793    77.0M   15.5G   609634756  303094956  101      155     26     77      458     
 SRR067784    83.3M   16.8G   660118460  328244560  101      155     26     77      458     
 SRR067785    81.6M   16.5G   646350512  321174108  101      155     26     77      458     
 SRR067792    83.8M   16.9G   663997828  330084304  101      155     26     77      458     
 SRR067577    46.3M   9.3G    367673108  183472948  101      155     26     77      458     
 SRR067579    46.0M   9.3G    365743380  182532676  101      155     26     77      458     
 SRR067578    46.5M   9.4G    369557476  184410788  101      155     26     77      458     
 
 #Jumping1 (mean insert size: 2283bp, SD 221), 101 bp read length
 SRR067771    81.5M   16.5G   644846296  320822716  101      2283    221    1620    2586    
 SRR067777    82.6M   16.7G   653163608  325232944  101      2283    221    1620    2586    
 SRR067781    82.1M   16.6G   649748720  323656576  101      2283    221    1620    2586    
 SRR067776    79.9M   16.1G   632590344  315165892  101      2283    221    1620    2586    
 
 #Jumping2 (mean insert size: 2803bp, SD 271), 101 bp read length
 SRR067773    93.1M   18.8G   736456192  366884512  101      2803    271    1990    3106    
 SRR067779    94.0M   19.0G   743564440  370214028  101      2803    271    1990    3106    
 SRR067778    97.3M   19.6G   767984324  381879652  101      2803    271    1990    3106    
 SRR067786    94.6M   19.1G   747631104  372002548  101      2803    271    1990    3106    
 
 #Fosmid1  (mean insert size: 35295bp, SD 2703), 76 bp read length
 SRR068214    13.1M   2.0G    104505420  52087176   76       35295   2703   27186   35523   36(trim 20bp at 5',20bp at 3')
 SRR068211    4.8M    736.9M  38612196   19252408   76       35295   2703   27186   35523   36(trim 20bp at 5',20bp at 3')
 
 #Fosmid2 (mean insert size: 35318bp, SD 2759),  101 bp read length
 SRR068335    67.4M   13.6G   533805860  265481252  101      35318   2759   27041   35621   61(trim 20bp at 5',20bp at 3')
  • Comments
    • Human chromosome 14. The chromosome may change, but this is a new data set with 100X coverage in 100bp and 76bp reads, just assembled by the Broad group using Allpaths-LG and Soap. We've downloaded the data and Todd is going to create a data set representing just chr 14, to make it feasible. We'll then try to assemble that data w/all 3 assemblers: CA, SOAP, Allpaths-LG.
  • Illumina chr14 reads (aligned with bowtie & corrected)
 /fs/szattic-asmg8/treangen/*fastq
 hard to align: bowtie -5 20 -3 20 -e 1000 ...
 jumping reads: only the ones aligned within coorect mean, stdev selected; these libraries usually have a high % of short inserts!!!

Assembly

Allpaths-lg

  • Read counts
                             orig       cor               cor(paired,all >64bp)
 chr14_fragment_12.fastq     36504800   35571477(97.44%)  34268444(10+bp ovl F/R)
 chr14_shortjump_12.fastq    22669408   11255320(49.64%)  11255320
 chr14_longjump_12.fastq     2405064    187398   (7.79%)  187398  
  • Assembly stats:
 .          elem  min    q1     q2     q3      max       mean     n50       sum       
 scf        418   96     131    256    1236    81646936  209781   81646936  87688255  
 scf10K+    17    10330  11780  26536  269876  81646936  5135452  81646936  87302692  
 ctg        4722  96     2342   9101   24174   240773    17887    36530       84461065  
  • Runtime 1104299.893u 126549.756s 18:50:05.80 1815.2% 0+0k 0+0io 8463pf+0w
 18hr 50min :                   multiprocessor
 1104299/(3600*24)=12.78 days : singleprocessor
  • Locations
 /scratch1/dpuiu/HTS/Homo_sapiens/Assembly/allpaths             # original
 /fs/szattic-asmg5/dpuiu/HTS/Homo_sapiens/Assembly/allpaths     # final contigs, scaff
 /fs/szattic-asmg5/dpuiu/HTS/Homo_sapiens/Assembly/allpathsCor  # corrected reads