Optimised DNA extraction and library preparation for minute arthropods: application to target enrichment in chalcid wasps used for biocontrol

Enriching subsets of the genome prior to sequencing allows focusing effort on regions that are relevant to answer specific questions. As experimental design can be adapted to sequence many samples simultaneously, using such approach also contributes to reduce cost. In the field of ecology and evolution, target enrichment is increasingly used for genotyping of plant and animal species or to better understand the evolutionary history of important lineages through the inference of statistically robust phylogenies. Limitations to routine target enrichment by research laboratories are both the complexity of current protocols and low input DNA quantity. Thus, working with tiny organisms such as micro-arthropods can be challenging. Here, we propose easy to set up optimisations for DNA extraction and library preparation prior to target enrichment. Prepared libraries were used to capture 1432 Ultra-Conserved Elements (UCEs) from microhymenoptera (Chalcidoidea), which are among the tiniest insects on Earth and the most commercialized worldwide for biological control purposes. Results show no correlation between input DNA quantities (1.8-250ng, 0.4 ng with an extra whole genome amplification step) and the number of sequenced UCEs. Phylogenetic inferences highlight the potential of UCEs to solve relationships within the families of chalcid wasps, which has not been achieved so far. The protocol (library preparation + target enrichment), allows processing 96 specimens in five working days, by a single person, without requiring the use of expensive robotic molecular biology platforms, which could help to generalize the use of target enrichment for minute specimens.

through the inference of statistically robust phylogenies. Limitations to routine target 26 enrichment by research laboratories are both the complexity of current protocols and low 27 input DNA quantity. Thus, working with tiny organisms such as micro-arthropods can be 28 challenging. Here, we propose easy to set up optimisations for DNA extraction and library 29 preparation prior to target enrichment. Prepared libraries were used to capture 1432 Ultra-30 Conserved Elements (UCEs) from microhymenoptera (Chalcidoidea), which are among the 31 tiniest insects on Earth and the most commercialized worldwide for biological control 32 purposes. Results show no correlation between input DNA quantities (1.8-250ng, 0.4 ng with 33 an extra whole genome amplification step) and the number of sequenced UCEs. Phylogenetic 34 inferences highlight the potential of UCEs to solve relationships within the families of chalcid 35 wasps, which has not been achieved so far. The protocol (library preparation + target 36 enrichment), allows processing 96 specimens in five working days, by a single person, 37 without requiring the use of expensive robotic molecular biology platforms, which could help 38 to generalize the use of target enrichment for minute specimens. 39

INTRODUCTION 45
Enriching subsets of the genome prior to sequencing (target enrichment, Mamanova et al. 46 2010) allows effort to be concentrated on genomic regions that are relevant to answer specific 47 research questions. Using this approach also contributes to reducing cost, as experimental 48 design can be adapted to sequence many samples simultaneously. In the fields of ecology and 49 evolution, target enrichment has been used for genotyping or phylogenomics of plant and 50 animal species (Gasc et al. 2016;Lemmon & Lemmon 2013), to characterize phenotypic 51 traits (e.g., Muraya et al. 2015) or to explore microbial ecosystems (Gasc & Peyret 2018). 52 53 However, routine target enrichment by research laboratories is limited both by the complexity 54 of current protocols, and by input DNA quantity that may be very low for some minute 55 species (e.g. micro-arthropods < 1mm) and /or old /rare (museum) specimens. Indeed, current 56 protocols (DNA extraction, library preparation, target-enrichment) are time consuming and 57 require handling expertise. They have been initially developed to work with large amounts of 58 input DNA (e.g., vertebrates or large/medium size insects; Faircloth et al. 2015;McCormack 59 et al. 2013) and include many purification steps that increase DNA loss. Working on hyper-60 diverse groups of microarthropods is challenging, as it requires one to perform the extraction 61 on i) a large number of specimens/species to be representative of the overall diversity of the 62 group, without the possibility of using pipetting robots that increase DNA loss, ii) single 63 individuals because species complexes are frequent (Al Khatib et al. 2014;Kenyon et al. 64 2015; Mottern & Heraty 2014), iii) the whole insect without destruction for vouchering and 65 often prior to species identification, iv) rare species that have been collected once and may be 66 represented in collection by a few specimens or only one specimen and, sometimes, v) old 67 and dry museum specimens used for species description (types). 68

69
In this study, we propose optimised protocols for DNA extraction and library preparation for 70 target enrichment purposes, as well as a custom pipeline to analyse the sequence data 71 obtained. We used these protocols and customised pipeline to capture and analyse Ultra 72 Conserved Elements (UCEs) in minute wasps, the chalcids (Insecta: Hymenoptera: 73 Chalcidoidea, Heraty et al. 2013;Noyes 2018), that are key components of terrestrial 74 ecosystems. Chalcids are key models for basic and applied research. With an estimated 75 diversity of more than 500,000 species these microhymenoptera have colonised almost all 76 extant terrestrial habitats. Many of them develop as parasitoids of arthropod eggs, larvae or 77 pupae. As such, they are both key regulators of the populations of many other arthropod 78 species in natural ecosystems and are increasingly used worldwide as biocontrol agents (e.g., 79 Consoli et al. 2010;Heraty 2009). A few of them, especially Nasonia (Pteromalidae) or 80 Trichogramma (Trichogrammatidae) species are also used as model systems to answer 81 challenging questions about sex determination, genetics of speciation, host-symbiont 82 interactions or behavioural ecology (e.g., Pinto et al. 1991;Stouthamer et al. 1990;Werren & 83 Loehlin 2009). Chalcidoidea has undergone a spectacular radiation resulting in a huge 84 diversity of morphologies and sizes (Gibson et al. 1999;Heraty et al. 2013), but are generally 85 small insects (< 2 mm long). Among them, Kikiki huna Huber (Mymaridae) at 158 µm long is 86 the smallest winged insect currently known and the wingless male of Dicopomorpha 87 echmepterygis Mockford at 139 µm is the smallest insect currently known (Huber & Noyes 88 2013). Notably, most species used for biological control, belonging mainly to five families 89 (Aphelinidae, Encyrtidae, Eulophidae, Mymaridae and Trichogrammatidae) are among the 90 tiniest wasps on earth (< 1mm). 91 92 Their small size, huge diversity and widespread morphological convergence make chalcid 93 wasps difficult to identify to species by non-expert taxonomists, which limits their use in 94 biological control. Attempts have been made to resolve the phylogeny of the whole 95 superfamily (Heraty et al. 2013;Munro et al. 2011) or a few families (Burks et al. 2011;96 Chen et al. 2004;Cruaud et al. 2012;Desjardins et al. 2007;Janšta et al. 2017;Owen et al. 97 2007) but none has succeeded. Indeed, the few markers that could be targeted with Sanger 98 sequencing were not informative enough to solve deeper relationships. A study based on 99 transcriptome data (3,239 single-copy genes) obtained from 37 species of chalcids and 11 100 outgroups also failed to solve relationships within the superfamily (Peters et al. 2018). As 101 only a representative sampling in both markers and taxa will allow one to draw accurate 102 conclusions on the history of this hyperdiverse group, target enrichment approaches appear 103 relevant. More specifically, targeting UCEs and their flanking regions that have been proven 104 useful to solve ancient and recent divergences (Faircloth et al. 2012;Smith et al. 2014) seems 105 pertinent. Indeed, a set of probes has been developed to target UCEs in Hymenoptera 106 (Faircloth et al. 2015). This set and an enriched one (Branstetter et al. 2017c) were 107 successfully used to solve the phylogeny of a few groups of ants, wasps and bees for which 108 the amount of DNA was not limiting (Blaimer et al. 2015;Blaimer et al. 2016a;Blaimer et al. 109 2016b;Bossert et al. 2017;Branstetter et al. 2017a;Branstetter et al. 2017b;Jesovnik et al. 110 (2015) using MYbaits kits (MYcroarray, Inc.). We followed manufacturer's protocol 170 (MYbaits, user manual version 3, http://www.mycroarray.com/pdf/MYbaits-manual-v3.pdf). 171 The hybridization reaction was run for 24h at 65°C. Post enrichment amplification was 172 performed on beads with the KAPA Hifi HotStart ReadyMix. The enriched libraries were 173 quantified with Qubit, an Agilent Bioanalizer and qPCR with the Library Quantification  Illumina/Universal from KAPA (KK4824). They were then pooled at equimolar ratio. Paired-175 end sequencing (2*300bp) was performed on an Illumina Miseq platform at UMR AGAP 176 (Montpellier, France) to get longer flanking regions and, as a consequence, more information 177 to differentiate closely related species. 178 179

Raw data cleaning 180
The analytical workflow is summarized in figure S1. In the next paragraph, UCEs, n=1432, additional files 5 and 6) using Geneious 8. 1.8 (Kearse et al. 2012) and contigs 194 were aligned to this set of reference UCEs using LASTZ Release 1.02.00 (Harris 2007). 195 Contigs that aligned with more than one reference UCE and different contigs that aligned with 196 the same reference UCE were filtered out using Geneious 8. 1.8. 197 198

Data analysis 199
UCEs for which sequences were available for more than 25% of the taxa were kept in the next 200 steps of the analysis. Alignments were performed with MAFFT v7.245 (Katoh & Standley 201 2013) (-linsi option). Ambiguously aligned blocks were removed using Gblock_0.91b with 202 relaxed constrains (-t=d -b2=b1 -b3=10 -b4=2 -b5=h) (Talavera & Castresana 2007). The 203 final data set was analysed using supermatrix approaches and coalescent-based summary 204 methods. Two gene tree reconciliation approaches were used: ASTRAL-III v5.6.1 (Zhang et 205 al. 2018), which computes the phylogeny that agrees with the largest number of quartet trees 206 induced by the set of input gene trees and ASTRID (Vachaspati & Warnow 2015) which 207 takes a set of gene trees, computes a distance matrix (ca sum of number of edges in the path 208 between two samples divided by the number of gene trees in which the two samples are 209 represented) and infers a phylogeny from this distance matrix. Following recommendations 210 for incomplete distance matrices, BioNJ was used to compute the phylogeny. Individual trees 211 were inferred from each UCE using raxmlHPC-PTHREADS-AVX (Stamatakis 2014) 212 (version 8.2.4; -f a -x 12345 -p 12345 -# 100 -m GTRGAMMA). ASTRAL and ASTRID 213 analyses were performed with 100 multi-locus bootstrapping (MLBS, site-only resampling 214 (Seo 2008)). Phylogenetic trees were estimated from the concatenate, unpartitioned data set 215 using Maximum Likelihood (ML) approaches as implemented in RAxML and IQTREE 216 v1.6.4 (Nguyen et al. 2015). For the RAxML analysis, a rapid bootstrap search (100 217 replicates) followed by a thorough ML search (-m GTRGAMMA) was performed. For the 218 IQTREE analysis, a ML search with the best-fit substitution model automatically selected was 219 performed with branch supports assessed with ultrafast bootstrap (Minh et al. 2013) and  aLRT test (Guindon et al. 2010 were recovered as monophyletic with high support. Aphelinids were split into three groups: 1) 254 a monophyletic Aphelininae + Eretmocerinae; 2) Coccophaginae; 3) Cales sp. (Calesinae). 255 The position of Cales was ambiguous. Cales was either recovered as sister to 256 Trichogrammatidae (RAxML, low support) or as a lineage distinct from all other chalcidoids 257 (all other analyses). Except for Mymaridae that was strongly placed as sister to all other 258 Chalcidoidea in all analyses, the tree backbone remained poorly resolved. Statistical support 259 was much higher within families. In all analyses Azotidae clustered with Signiphoridae, with 260 strong support. 261 262 DISCUSSION 263 To our knowledge this study is the second after Sproul and Maddison (2017) to demonstrate 264 success in library preparation from such low input using commercial kits, and the first to 265 report successful sequencing of >1000 low copy genes in 96 specimens in parallel, from such 266 low input and processing time. Our optimisations differ from what was proposed by Sproul 267 and Maddison (2017). First, we tried to optimize DNA extraction itself by using overnight 268 lysis with gentle mixing to preserve fragile specimens, heated elution buffer and increased 269 incubation time before elution. Second, instead of increasing the number of time-consuming 270 purification steps we decreased them. It is noteworthy that in this library only two historical 271 specimens were included. This may have masked challenges posed by adapter dimers (Burrell 272 et al. 2015;Sproul & Maddison 2017;Tin et al. 2014) that led Sproul and Maddison (2017) to 273 add a second bead clean-up prior to library amplification. However, we have already used this 274 protocol on hundreds of chalcid and moth species, including historical specimens that were 275 processed the same way as fresh ones and we never had such an issue. Finally, instead of 276 increasing the number of amplification cycles prior to target enrichment, we used a new 277 generation mastermix including a hot start, processive and high-fidelity polymerase 278 (NEBNext Ultra II Q5 Master Mix). Our protocol also allows one to back-up DNA at several 279 steps that allows for multiple attempts without delay in case the first attempt fails. Finally, 280 sequencing was performed on a MiSeq to get longer flanking regions and, as a consequence, 281 more information to differentiate closely related species. 282

283
The protocol was successfully used on minute chalcid wasps widely used for biological 284 control purposes. Up to 1165 valid UCEs were captured from 25ng DNA (median amount of 285 DNA used for this study). No correlation was observed between the quantity of DNA used for 286 library preparation and the number of captured UCEs. The average number of captured UCEs 287 validated by our quality control workflow was 687, and 685 valid UCEs were captured from a 288 tiny aphelinid (1.8ng of input DNA). The number of UCEs obtained per individual appears to 289 drop within the basal clades (i.e., Mymaridae and Trichogrammatidae, Figure S2), a result 290 probably linked to the relatively long branches observed in these groups, and that could 291 reduce the efficiency of the probes that were designed from the genome of Nasonia (Faircloth 292 et al. 2015). Trees were well resolved at the family level, with high statistical support, 293 showing the potential of UCEs to solve long-standing taxonomic issues. However, the tree 294 backbone remained unresolved, a pattern that confirms the rapid diversification of the group 295 (Heraty et al. 2013;Peters et al. 2018). Understanding the evolutionary history of the group 296 was not the purpose of this paper. Indeed, only a representative sampling in both markers and 297 taxa as well as cutting-edge data analysis will allow drawing accurate conclusions. By 298 providing suitable tools for a fast, easy and affordable acquisition of data, this paper is a first 299 step. 300 301 Interestingly, as it has been shown previously on RADseq data (Cruaud et al. in press), WGA 302 does not seem to bias the results even when input DNA used for the WGA is below the 303 recommendation on the manufacturer kit (here 0.4 and 4 ng when the Genomiphi kit V2 304 requires 10 ng). This further increases the possibilities opened by our protocol as input DNA 305 as low as 0.2-0.3 ng may be used when an extra WGA step is included after DNA extraction 306 (Cruaud et al. in press). It is noteworthy that reducing the amount of DNA required for library 307 preparation allows one to use extracted DNA for different approaches in parallel (RADseq,308 amplicon,Shotgun etc.). This also allows one to send DNA back to museums from which 309 specimens were borrowed, for archival purposes. We definitely agree with Sproul and 310 Maddison (2017) who emphasize how important it is not to waste DNA obtained from 311 irreplaceable specimens (whether fresh or historical). It is even more important to capitalize 312 on existing collections as collecting samples for large-scale studies may be more and more 313 difficult, given that many countries have imposed restrictive access regulations, even to 314 academic researchers, to reduce the risk of supposed biopiracy (Divakaran Prathapan et al. 315 2018). 316 317 All the elements discussed above indicate that this protocol may be of great help to 318 reconstruct phylogenetic hypotheses in multiple groups of tiny arthropods, e.g., springtails, 319 Al Khatib F, Fusu L, Cruaud A, et al. (2014)  Biology 61, 717-726. 410 Gagnaire PA, Normandeau E, Pavey SA, Bernatchez L (2013) Mapping phenotypic, 411 expression and transmission ratio distortion QTL using RAD markers in the Lake 412 Whitefish (Coregonus clupeaformis). Molecular Ecology 22, 3036-3048. 413 Gasc C, Peyret P (2018) Hybridization capture reveals microbial diversity missed using 414 current profiling methods. Microbiome 6, 61. 415 Gasc C, Peyretaillade E, Peyret P (2016) Sequence capture by hybridization to explore 416 modern and ancient genomic diversity in model and nonmodel organisms. Nucleic Acids 417 Research 44, 4504-4518. 418 Gibson GAP, Heraty JM, Woolley JB (1999) Phylogenetics and classification of Chalcidoidea 419 and Mymarommatoidea -a review of current concepts (Hymenoptera, Apocripta). 420 Zoologica Scripta 28, 87-124. 421 Gibson GAP, Read J, Huber JT (2007)    Black squares indicate node supported with RAxML BP = 100, IQTREE aLRT = 100 / BP = 100 and ASTRAL/ASTRID BP > 75. Grey square indicates node with RAxML BP = 100, IQTREE aLRT = 100 / BP = 100 and ASTRAL BP > 75. White squares indicate nodes with RAxML BP > 95 and IQTREE aLRT > 80 / BP > 95. Figure S3: IQTREE tree obtained from the analysis of the 25%-complete data set (page 138) SH-aLRT / UFboot values are indicated at nodes. The DNA quantity used to build the library as well as the number of UCEs analysed for each sample is given in brackets.

Figure S4
ASTRAL tree obtained from the analysis of the 25%-complete data set (page 139) Bootstrap supports (site-only resampling) are indicated at nodes. The DNA quantity used to build the library as well as the number of UCEs analysed for each sample is given in brackets.

Figure S5
ASTRID tree obtained from the analysis of the 25%-complete data set (page 140) Bootstrap supports (site-only resampling) are indicated at nodes. The DNA quantity used to build the library as well as the number of UCEs analysed for each sample is given in brackets.

Additional file 1: Optimized protocol for DNA extraction
Optimisations were done to the DNeasy Blood & Tissue Kit (250) protocol (Qiagen)
• Equilibrate frozen specimens to room temperature before processing.
• Dry specimens on clean paper towel to remove EtOH. Place specimen in tube. Clean forceps to avoid contamination between 2 different specimens: bleach, then water, then dry with paper towel.
• Add 200µl buffer ATL (Qiagen) to each tube. (NB DNeasy® Blood & Tissue Handbook uses 180µl; we use 200µl to improve diffusion of the DNA from the specimen) • Add 20µl proteinase K to each tube (or prepare a master mix of buffer ATL + proteinase K and add master mix to tubes.) • Vortex + Pulse spin. (NB if the specimens are fragile, i.e. old and dry mounted collection specimens, dispense the master mix of buffer ATL + proteinase K to each tube and then add specimens; or dispense specimens to tubes and then add the master mix. Do not vortex at all to avoid damage to specimens and place the tubes directly in the thermomixer after ensuring that the specimens are in the liquid.) • Place tubes in thermomixer (Eppendorf) at 56°C and 300 rpm overnight.
• Vortex + pulse spin. (NB if working with fragile specimens do not vortex. Pulse spin, transfer all the liquid to a new labelled tube leaving the specimen behind and then proceed to the next step.) (DNeasy® Blood & Tissue Handbook, p. 15, recommends using carrier DNA or RNA for samples containing less than 5 µg of DNA. Though this was not used for this publication, poly (A) RNA (2 µl of 2 µg/µl) is added to the lysate before the next step by one of us (LF) for routine DNA purification from small or old specimens. This should not interfere with library preparation as it was used for example by Sproul & Maddison (2017) as recommended in the QIAamp DNA Micro Kit by Qiagen.) • Add 200µl buffer AL to each tube (kit Qiagen). (DNeasy® Blood & Tissue Handbook,p. 19,29 recommends adding RNase A if RNA-free genomic DNA is required. We do not use it because the copurified RNA is not interfering with target enrichment anyway and because RNase A also degrades DNA; see Donà & Houseley (2015).) A white precipitate may appear that will dissolve during incubation at 70°C. • Vortex + pulse spin.
• Incubate tubes in thermomixer at 70°C and 300 rpm for 10 minutes.
• During that time, prepare DNeasy mini spin Qiagen columns and 1.5 ml DNA LoBind Eppendorf tubes (label tubes with sample codes). (NB LoBind tubes minimize DNA loss during storage by reducing sample-to-surface binding.) • Aliquot buffer AE in Eppendorf tubes and heat at 56°C in thermomixer (heated buffer will improve the release of DNA from the resin.) • Add 200µl absolute ethanol (not provided with the Qiagen kit) to each tube. (NB only high purity ethanol should be used; cheap ethanol may contain traces of other chemicals that will interfere with the solubilisation of the DNA from the membrane in the last step.) • Vortex + Pulse spin.
• Pipette the liquid (including precipitate) from each tube and transfer into a DNeasy mini spin Qiagen column placed in a 2 ml collection tube (Qiagen kit).
• Centrifuge columns + collection tubes at 6000 x g (8000 rpm) for 1 minute; discard the flow-through and collection tubes. Keep the columns.
• Place spin columns in new collection tubes.
• Centrifuge at 6000 x g (8000rpm) for 1 minute; discard the flow-through and collection tubes. Keep the columns.
• Place spin columns in new collection tubes.
• Centrifuge at 20,000 x g (14,000 rpm) for 3 minutes to dry the columns.
• Rotate columns to 180 degrees in their collection tubes and centrifuge again at 20000 xg (14000rpm) for 3 minutes (this will make sure the column is well dried).
• Place dried spin columns in 1.5ml LoBind tubes that you previously labelled.
• Add 50µl of heated AE buffer. Deposit buffer right in the middle of the resin.
• Rotate columns to 180 degrees and centrifuge again at 6000 x g (8000 rpm) for 1 minute.
• Remove columns + tubes from the centrifuge.
• Add again 50µl of heated buffer right in the middle of the resin.
• Incubate 15 min at room temperature.
• Centrifuge at 6000 x g (8000 rpm) for 1 minute at room temperature.
• Rotate columns to 180 degrees and centrifuge again at 6000 x g (8000 rpm) for 1 minute.
• DNA (ca 2 x 50µl) is ready for use.
• Add distilled water to the extracted specimens and incubate at room temperature for 30 min.
(NB water is used to eliminate the residual salts left by the buffers. Otherwise they usually crystallise on the specimen upon storage in ethanol.) • Remove water and replace with ethanol. Keep specimens in the dark in a freezer until mounting. (NB if the specimens are stored at room temperature under ambient light they will become sun-bleached very quickly, much quicker than unextracted specimens.) • Use a critical point dryer, hexamethyldisilazane, acetone or amyl acetate to dry the specimens before mounting. Do not air dry as they will collapse.

DNA normalization and shearing
Bioruptor ® Pico, sonication bath-based rotor 12 samples in parallel.
Transfer 100µl of input DNA in tubes adapted to the Bioruptor ® Pico that will be used for DNA shearing (Diagenode tubes 0.5ml). [now we shear 30ng DNA in routine but you can use as little as 1.8 ng, see main text] Set up the following program for DNA shearing to obtain a mean fragment size of ca 400pb: 15sec ON / 90sec OFF for 8 cycles.
You can STOP here for up to 1 week (store tubes in a freezer at -20°C).
The NEBNext Ultra II DNA Library prep kit for Illumina by NEB will be used for End repair, A tailing, Adapter ligation and PCR enrichment of adaptor-ligated DNA.

End repair + A tailing
To keep a backup, only 50 µl of the sheared DNA is used [you may want to re-concentrate DNA on beads in 50 µl when you have a small amount i.e. < 5 ng] Transfer 50 µl of sheared DNA into a strip / plate.

Adapter ligation and cleanup
See additional file 3 for adapter sequences and hybridization.
For 96 samples use the following tagging scheme: Thaw NEBNext Ultra II Ligation master mix. Light vortex + Pulse spin.

Clean-up adaptor ligated DNA (remove dimers of adapters):
Purification was conducted with the Agencourt AMPure XP Purification system (Beckman Coulter) and a ALPAQUA magnet plate.
Pull Ampure beads out of fridge. Make sure they are at room temperature (RT) before use.
In our experience, a clean-up step of the librairies is not necessary.
Pool 100 ng of each library (you want a final quantity of 100-500ng DNA per pool after concentration to fit with the requirements of the MYbaits® protocol.) On beads Concentration of DNA (Agencourt AMPure XP Purification system (Beckman Coulter)). Concentration of DNA is required, as MyBaits require 7 µl of starting material.
Remove Ampure beads from fridge. Make sure they are at room temperature (RT) before use. Qubit quantify 1 µl of each library (dilution 1:2).
Check library profile on an Agilent Bioanalyzer with a High Sensitivity DNA Analysis Kit (load 1µl of library dilution 1:2).

PCR Enrichment of captured libraries and final clean-up
At this step of the process we have obtained libraries and Streptavidin beads mixed in 30 µl of 10mM Tris-Cl, 0.05% TWEEN®-20 solution (pH 8.0 -8.5).
Here we want to release the bead-bound UCE-enriched library from the baits and amplif the resulting fragments.
PCR enrichment of the captured fragment is performed on beads with the following mix. We performed 2 PCR reactions for each pool of samples. -Equimolar pooling of the 6 librairies.