We did assembly by SPAdes with Dataset 4 raw data first, and then used different subreads depths of Dataset 5 and Dataset 9 to scaffold by SSPACE-longread.

We arbitrary chose 1-4 SMRT cells:
One single SMRT cell: m120208_071634
Two SMRT cells: m120228_210845 + m120208_122534
Three SMRT cells: m120228_115504 + m120228_152936 + m120228_100807
Four SMRT cells: m120228_171636 + m120228_223624 + m120228_100807 + m120228_190630

spades.py -1 reads_1.fastq -2 reads_2.fastq -o output
SSPACE-LongRead.pl -c contig.fasta -p filter_subreads.fasta -b output

Evaluation

We have evaluated the assemblies with QUAST 2.3(reference genome NC_000913 and Ec_gene_list). more detail

Statistics without reference	Miseq_only_sspace	Miseq_1cell_sspace	Miseq_2cell_sspace	Miseq_3cell_sspace	Miseq_4cell_sspace	Miseq_17cell_sspace	Miseq_d9_sspace
# contigs	86	15	18	16	15	17	14
Largest contig	285889	2497845	1260980	2501081	3194637	1954649	3392211
Total length	4577132	4632009	4633058	4636174	4638657	4633857	4632677
N50	139882	2497845	1238868	2501081	3194637	1238635	3392211
Misassemblies
# misassemblies	2	9	10	10	10	8	8
Misassembled contigs length	215581	3193893	3244631	2705788	3657574	3243566	4050696
Mismatches
# mismatches per 100 kbp	3.02	7.15	6.2	6.9	7.17	6.05	6.43
# indels per 100 kbp	0.46	1.06	1	1.32	1.27	0.95	1.06
# N's per 100 kbp	0	97.89	67.67	77.33	77.37	91.03	123.88
Genome statistics
Genome fraction (%)	98.451	99.498	99.483	99.664	99.748	99.432	99.587
Duplication ratio	1.001	1.002	1.002	1.002	1.002	1.003	1.001
# genes	4399 + 32 part	4467 + 14 part	4465 + 13 part	4476 + 11 part	4477 + 11 part	4467 + 11 part	4470 + 13 part
NGA50	133059	571664	425173	852639	1039467	1039472	1039654
Running Time	2hr 28m	2hr 32m	2hr 37m	2hr 43m	2hr 47m	3hr 43m	3hr 11m

Misassemblies for Adobe reader.