goal: NIAID sent MiSeq estimates for Pt3 1995 there should be 20
(s1:s20), check against our results.

================================

SUMMARY:


- We sequenced four mixes of 3, 5, 6, and 6 SGA shorter LID HIV
  genomes from Pt3/1995 blinded to the truth.

- These results use the de-novo Long Amplicon Analysis (LAA)
  push-button protocol publicly available in our SMRTportal software.

- We saw 100% concordance across the entire MiSeq estimate (except ~3
  bases at the begining in some) with the PacBio estimate adding ~30
  bases to the begining and ending of the MiSeq estimate.

- We correctly de-novo estimated the genomes even though one sample
  contained multiple genomes and several samples contained identical
  genomes.

- This "out-of-the-box" solution warrants unblinding the remaining Pt3
  sequences (our estimates at the end of this page). If those look
  good, then testing on the 20-mix might give perfect results.

================================

NIAID unblinded the Pt3 1995 MiSeq estimates (Friday 2014-6-6)

  -rw-r--r-- 1 mbrown domain_users     71570 2014-06-06 15:31 pt3_20_short_frag_1995_sample_contigs.fa
  -rw-r--r-- 1 mbrown domain_users     11233 2014-06-06 15:31 pt3_20_short_frag_1995_sample_info.xlsx

Here is the spreadsheet than Yan joined with our data:
-=-=-=-=





mixNameLIBSampleNameHiromiSampleNameHiromiEstSizeAgilentLengthMiSeqDelLengthAproxDelHXB2CGClusterNote


Pt3-95-3mix-A-short S1_hiromi_3rd_vicunaCon(2536) Pt3_s#19 1828      heterozygous Not a SGA, i.e. containing multiple HIV genomes 


Pt3-95-3mix-A-short S8_vicuna_consensus(2884) Pt3_#s2 282527842788&3377 2602..5390 & 5401..8778 Unique 


Pt3-95-3mix-A-short S2_hiromi_3rd_vicunaCon(2261) Pt3_s#20 2252216167201082..7801 cross p17 and gp41 Unique 6720 bp deletion causes frameshift in gp41 





Pt3-95-3mix-B-short S7_vicuna_consensus(3165) Pt3_#s1 3132306558703682..9552 Unique 


Pt3-95-3mix-B-short S16_vicuna_consensus(3422) Pt3_#s10 3541332256093279..8888 D All identical, but with slightly different starting/ending position


Pt3-95-3mix-B-short S17_vicuna_consensus(3425) Pt3_#s11 3533332556093279..8888 D All identical, but with slightly different starting/ending position


Pt3-95-3mix-B-short S9_vicuna_consensus(3416) Pt3_#s3 3542331656093279..8888 D All identical, but with slightly different starting/ending position


Pt3-95-3mix-B-short S14_vicuna_consensus(3415) Pt3_#s8 3483331556093279..8888 D All identical, but with slightly different starting/ending position





Pt3-95-3mix-C-short S18_vicuna_consensus(3418) Pt3_#s12 3593331856093279..8888 D All identical, but with slightly different starting/ending position


Pt3-95-3mix-C-short S19_vicuna_consensus(3429) Pt3_#s13 3709332956093279..8888 D All identical, but with slightly different starting/ending position


Pt3-95-3mix-C-short S20_vicuna_consensus(3439) Pt3_#s14 3665333956093279..8888 D All identical, but with slightly different starting/ending position


Pt3-95-3mix-C-short S22_vicuna_consensus(3431) Pt3_#s16 3670333156093279..8888 D All identical, but with slightly different starting/ending position


Pt3-95-3mix-C-short S24_vicuna_consensus(3452) Pt3_#s18 3660335256093279..8888 D All identical, but with slightly different starting/ending position


Pt3-95-3mix-C-short S10_vicuna_consensus(3418) Pt3_#s4 3634331856093279..8888 D All identical, but with slightly different starting/ending position





Pt3-95-3mix-D-short S21_vicuna_consensus(5092) Pt3_#s15 53974992320&36244857..5177 & 5464..9088 Unique Also has 285 bp invertion (4466..4181 in consensus or 5178..5463 in HXB2) 


Pt3-95-3mix-D-short S23_vicuna_consensus(5050) Pt3_#s17 5483495039605621..9581 Unique 


Pt3-95-3mix-D-short S11_vicuna_consensus(4279) Pt3_#s5 4302417947294354..9083 Unique 


Pt3-95-3mix-D-short S12_vicuna_consensus(4742) Pt3_#s6 4879464242834781..9064 Unique 


Pt3-95-3mix-D-short S13_vicuna_consensus(4770) Pt3_#s7 4919467042844775..9059 Unique 


Pt3-95-3mix-D-short S15_vicuna_consensus(5237) Pt3_#s9 5481513737645640..9404 Unique 




-=-=-=-=

We sequenced four mixes of 3, 5, 6, and 6 SGA shorter LID HIV genomes
from Pt3/1995 blinded to the truth.

Here's the number of mixed SGA samples and number unique genomes per
mix based on MiSeq:

run numSGAs numUniqMiSeqGenomes
--- ------- -------------------
A   3       4+
B   5       2
C   6       1
D   6       6
--- ------- -------------------

Note that the number of SGA samples and the number of unique genomes
differ.

Mixes E,F,G are not from 1995 and are still blinded, though they were
run and analyzed.

================================================================


#### LAA RESULTS


See end of page for definitions of terms in result tables.

================================
----Mix A:

RESULTS:
mixA-amplicon_analysis_summary.csv
mixA-amplicon_analysis.fasta

mymerge[mymerge$TotalCov>200,]
                             FastaName BarcodeName CoarseCluster  Phase
1 Barcode0_Cluster0_Phase0_NumReads298           0      Cluster0 Phase0
2 Barcode0_Cluster0_Phase1_NumReads202           0      Cluster0 Phase1
3 Barcode0_Cluster1_Phase0_NumReads500           0      Cluster1 Phase0
4 Barcode0_Cluster2_Phase0_NumReads328           0      Cluster2 Phase0
  TotalCoverage SequenceLength PredictedAccuracy ConsensusConverged
1           298           2238         0.9995531               True
2           202           2234         0.9975511              False
3           500           2818         1.0000000               True
4           328           1628         1.0000000               True
  NoiseSequence IsChimera ChimeraScore ParentSequenceA ParentSequenceB
1         False     False          NaN              NA              NA
2         False     False          NaN              NA              NA
3         False     False          NaN              NA              NA
4         False     False          NaN              NA              NA
  CrossoverPosition                                       qName
1                -1 Barcode0_Cluster0_Phase0_NumReads298/0_2238
2                -1 Barcode0_Cluster0_Phase1_NumReads202/0_2234
3                -1 Barcode0_Cluster1_Phase0_NumReads500/0_2818
4                -1 Barcode0_Cluster2_Phase0_NumReads328/0_1628
                                            tName qStrand tStrand  score
1 S2_hiromi_full_length_3rd_vicunaConsensus(2261)       0       0 -10805
2                      S13_vicuna_consensus(4770)       0       1  -2681
3                       S8_vicuna_consensus(2884)       0       1 -13905
4                       S7_vicuna_consensus(3165)       0       0  -4507
  percentSimilarity tStart tEnd tLength qStart qEnd qLength nCells
1          100.0000      0 2161    2161     36 2197    2238  45349
2           98.2079   4112 4670    4670   1661 2218    2234  11682
3          100.0000      0 2781    2784     31 2812    2818  58369
4           96.9792      3  937    3065      6  966    1628  20865

We get 2 100% about full-length and extra on ends (S2 and S8). S1 was
not an SGA and we get two imperfect hits to (S13 and S7). If S1 was
really two genomes then we are correct.

NOTE: Accuracy of hit in percentSimilarity. Full-length because on
target: tStart~=0 and tEnd~=tLength. Extra on ends because on query:
qStart > 0 and qEnd < qLength

================================
----Mix B:

RESULTS:
mixB-amplicon_analysis_summary.csv
mixB-amplicon_analysis.fasta

mymerge[mymerge$TotalCov>200,]
                             FastaName BarcodeName CoarseCluster  Phase
1 Barcode0_Cluster0_Phase0_NumReads500           0      Cluster0 Phase0
2 Barcode0_Cluster1_Phase0_NumReads500           0      Cluster1 Phase0
  TotalCoverage SequenceLength PredictedAccuracy ConsensusConverged
1           500           3373                 1               True
2           500           3095                 1               True
  NoiseSequence IsChimera ChimeraScore ParentSequenceA ParentSequenceB
1         False     False          NaN              NA              NA
2         False     False          NaN              NA              NA
  CrossoverPosition                                       qName
1                -1 Barcode0_Cluster0_Phase0_NumReads500/0_3373
2                -1 Barcode0_Cluster1_Phase0_NumReads500/0_3095
                       tName qStrand tStrand  score percentSimilarity tStart
1 S24_vicuna_consensus(3452)       0       0 -16745               100      0
2  S7_vicuna_consensus(3165)       0       0 -15310               100      3
  tEnd tLength qStart qEnd qLength nCells
1 3349    3352     18 3367    3373  70297
2 3065    3065      6 3068    3095  64270

We get 2 100% about full-length and extra on ends (S7 and
S24). Correct! S24 is the best hit among the "all identical".

================================
----Mix C:

RESULTS:
mixC-amplicon_analysis_summary.csv
mixC-amplicon_analysis.fasta

mymerge[mymerge$TotalCov>200,]
                             FastaName BarcodeName CoarseCluster  Phase
1 Barcode0_Cluster0_Phase0_NumReads500           0      Cluster0 Phase0
  TotalCoverage SequenceLength PredictedAccuracy ConsensusConverged
1           500           3372                 1               True
  NoiseSequence IsChimera ChimeraScore ParentSequenceA ParentSequenceB
1         False     False          NaN              NA              NA
  CrossoverPosition                                       qName
1                -1 Barcode0_Cluster0_Phase0_NumReads500/0_3372
                       tName qStrand tStrand  score percentSimilarity tStart
1 S24_vicuna_consensus(3452)       0       0 -16745               100      0
  tEnd tLength qStart qEnd qLength nCells
1 3349    3352     17 3366    3372  70297

We get 1 100% about full-length and extra on ends (S24). Correct! S24
is the best hit among the "all identical". This is given six input SGA
tubes.

================================
----Mix D:

RESULTS:
mixD-amplicon_analysis_summary.csv
mixD-amplicon_analysis.fasta

> mymerge[mymerge$TotalCov>200,]
                              FastaName BarcodeName CoarseCluster  Phase
1  Barcode0_Cluster0_Phase0_NumReads500           0      Cluster0 Phase0
16 Barcode0_Cluster2_Phase0_NumReads500           0      Cluster2 Phase0
17 Barcode0_Cluster3_Phase0_NumReads422           0      Cluster3 Phase0
18 Barcode0_Cluster4_Phase0_NumReads341           0      Cluster4 Phase0
19 Barcode0_Cluster5_Phase0_NumReads285           0      Cluster5 Phase0
20 Barcode0_Cluster6_Phase0_NumReads234           0      Cluster6 Phase0
   TotalCoverage SequenceLength PredictedAccuracy ConsensusConverged
1            500           4234                 1               True
16           500           4695                 1               True
17           422           5201                 1               True
18           341           4698                 1               True
19           285           5019                 1               True
20           234           5012                 1               True
   NoiseSequence IsChimera ChimeraScore ParentSequenceA ParentSequenceB
1          False     False          NaN                                
16         False     False          NaN                                
17         False     False          NaN                                
18         False     False          NaN                                
19         False     False          NaN                                
20         False     False          NaN                                
   CrossoverPosition                                       qName
1                 -1 Barcode0_Cluster0_Phase0_NumReads500/0_4234
16                -1 Barcode0_Cluster2_Phase0_NumReads500/0_4695
17                -1 Barcode0_Cluster3_Phase0_NumReads422/0_5201
18                -1 Barcode0_Cluster4_Phase0_NumReads341/0_4698
19                -1 Barcode0_Cluster5_Phase0_NumReads285/0_5019
20                -1 Barcode0_Cluster6_Phase0_NumReads234/0_5012
                        tName qStrand tStrand  score percentSimilarity tStart
1  S11_vicuna_consensus(4279)       0       1 -20895               100      0
16 S13_vicuna_consensus(4770)       0       1 -23350               100      0
17 S15_vicuna_consensus(5237)       0       1 -25685               100      0
18 S12_vicuna_consensus(4742)       0       1 -23210               100      0
19 S23_vicuna_consensus(5050)       0       1 -24750               100      0
20 S21_vicuna_consensus(5092)       0       0 -24945               100      3
   tEnd tLength qStart qEnd qLength nCells
1  4179    4179     23 4202    4234  87727
16 4670    4670      9 4679    4695  98038
17 5137    5137     34 5171    5201 107845
18 4642    4642     34 4676    4698  97450
19 4950    4950     33 4983    5019 103918
20 4992    4992      7 4996    5012 104737

We got all six of them right with coverage 200 cutoff! CORRECT!

All of them full length (except 3 bases in S21) with and extra 30
bases or so at the ends, so we extend the MiSeq estimate!

================================================================================================


Results for the Pt3 2001 still blinded runs.


Estimates with coverages>200 should be examined

----Mix E (2001: s29, s30, s31)

RESULTS:
mixE-amplicon_analysis_summary.csv
mixE-amplicon_analysis.fasta

----Mix F (2001: s21, s23, s27, s32)

RESULTS:
mixF-amplicon_analysis_summary.csv
mixF-amplicon_analysis.fasta

----Mix G (2001: s22, s24, s25, s26, s28)

RESULTS:
mixG-amplicon_analysis_summary.csv
mixG-amplicon_analysis.fasta

================================================================================================


Definitions for result tables:


- FastaName:	name of the consensus sequence
- BarcodeName:	always 0 in this case, if using barcodes then the id
- CoarseCluster: how the algorithm split the reads
- Phase:	how the algorithm split the CoarseCluster
- TotalCoverage: number of reads that went into consensus
- SequenceLength: the length of the estimate consensus
- PredictedAccuray: the algorithm's de-novo estimate of the consensus accuracy
- ConsensusConverged: whether the algorithm converged confidently to an answer
- NoiseSequence: is the consensus estimate likely to be noise
- IsChimera: is the consensus a cross-over of other sequences? Important in PCRing related species
- ChimeraScore ParentSequenceA ParentSequenceB CrossoverPosition: information if IsChimera=True
- qName: queryID in alignment
- tName: targetID in alignment
- qStrand, tStrand: direction of query and target in alignment
- score: alignment score
- percentSimilarity: the similarity in the aligned region
- tStart, tEnd: where the alignment started and ended in the target sequence
- tLength: total length of the target sequence
- qStart, qEnd: where the alignment started and ended in the query sequence
- qLength: total length of the query sequence
- nCells: Number of ZMWs used??? I'll have to ask.

Example:
                             FastaName BarcodeName CoarseCluster  Phase
1 Barcode0_Cluster0_Phase0_NumReads500           0      Cluster0 Phase0
  TotalCoverage SequenceLength PredictedAccuracy ConsensusConverged
1           500           3372                 1               True
  NoiseSequence IsChimera ChimeraScore ParentSequenceA ParentSequenceB
1         False     False          NaN              NA              NA
  CrossoverPosition                                       qName
1                -1 Barcode0_Cluster0_Phase0_NumReads500/0_3372
                       tName qStrand tStrand  score percentSimilarity tStart
1 S24_vicuna_consensus(3452)       0       0 -16745               100      0
  tEnd tLength qStart qEnd qLength nCells
1 3349    3352     17 3366    3372  70297