Low-abundance Sequences Research Articles

THAPBI PICT-a fast, cautious, and accurate metabarcoding analysis pipeline.

THAPBI PICT is an open source software pipeline for metabarcoding analysis of Illumina paired-end reads, including cases of multiplexing where more than one amplicon is amplified per DNA sample. Initially a Phytophthora ITS1 Classification Tool (PICT), we demonstrate using worked examples with our own and public data sets how, with appropriate primer settings and a custom database, it can be applied to other amplicons and organisms, and used for reanalysis of existing datasets. The core dataflow of the implementation is (i) data reduction to unique marker sequences, often called amplicon sequence variants (ASVs), (ii) dynamic thresholds for discarding low abundance sequences to remove noise and artifacts (rather than error correction by default), before (iii) classification using a curated reference database. The default classifier assigns a label to each query sequence based on a database match that is either perfect, or a single base pair edit away (substitution, deletion or insertion). Abundance thresholds for inclusion can be set by the user or automatically using per-batch negative or synthetic control samples. Output is designed for practical interpretation by non-specialists and includes a read report (ASVs with classification and counts per sample), sample report (samples with counts per species classification), and a topological graph of ASVs as nodes with short edit distances as edges. Source code available from https://github.com/peterjc/thapbi-pict/ with documentation including installation instructions.

FLASH-TB: an Application of Next-Generation CRISPR to Detect Drug Resistant Tuberculosis from Direct Sputum

Offering patients with tuberculosis (TB) an optimal and timely treatment regimen depends on the rapid detection of Mycobacterium tuberculosis (Mtb) drug resistance from clinical samples. Finding Low Abundance Sequences by Hybridization (FLASH) is a technique that harnesses the efficiency, specificity, and flexibility of the Cas9 enzyme to enrich targeted sequences. Here, we used FLASH to amplify 52 candidate genes probably associated with resistance to first- and second-line drugs in the Mtb reference strain (H37Rv), then detect drug resistance mutations in cultured Mtb isolates, and in sputum samples. 92% of H37Rv reads mapped to Mtb targets, with 97.8% of target regions covered at a depth ≥ 10X. Among cultured isolates, FLASH-TB detected the same 17 drug resistance mutations as whole genome sequencing (WGS) did, but with much greater depth. Among the 16 sputum samples, FLASH-TB increased recovery of Mtb DNA compared with WGS (from 1.4% [IQR 0.5-7.5] to 33% [IQR 4.6-66.3]) and average depth reads of targets (from 6.3 [IQR 3.8-10.5] to 1991 [IQR 254.4-3623.7]). FLASH-TB identified Mtb complex in all 16 samples based on IS1081 and IS6110 copies. Drug resistance predictions for 15/16 (93.7%) clinical samples were highly concordant with phenotypic DST for isoniazid, rifampicin, amikacin, and kanamycin [15/15 (100%)], ethambutol [12/15 (80%)] and moxifloxacin [14/15 (93.3%)]. These results highlighted the potential of FLASH-TB for detecting Mtb drug resistance from sputum samples.

Evidence-based strategies to navigate complexity in dietary DNA metabarcoding: A reply.

It is clearly beneficial to eliminate low-abundance sequences that arise in error during dietary DNA metabarcoding studies, but to purge all low-abundance sequences is to risk eliminating real sequences and complicating ecological analyses. Our prior literature review noted that DNA sequence relative read abundance (RRA) thresholds can help ameliorate false-positive taxon occurrences, but that historical emphasis on this utility has fostered uncertainty about the associated risk of inflating the false-negative rate (Littleford-Colquhoun et al.,2022). To address this, we combined a simulation study and an empirical data set to both illustrate the issue and provide blueprints for simulation studies and sensitivity analyses that can help investigators avoid overcorrecting and thereby bolster confidence in ecological inferences. Awareness of both the costs and the benefits of abundance-filtering is needed because accurately characterizing dietary distributions can be critically important for understanding animal diets, nutrition and trophic networks. Highlighting the need to raise awareness, a critique of our paper emphasized the misleading notion that "false positive interactions between species can present fundamentally incorrect network structures in network ecology, whereas false negatives will provide a correct but incomplete version of the network" (Tercel & Cuff,2022). Asserting that the reliability of results will be eroded by false positives but resilient to the omission of true positives is risky and runs counter to evidence. Unfortunately, abundance-filtering methods can introduce false negatives at higher rates than they eliminate false positives and thereby undermine the analysis of otherwise reliable sequencing data. Overcorrecting can qualitatively alter and ultimately undermine ecological interpretations.

Culture-Free Phylogenetic Analysis of Legionella pneumophila Using Targeted CRISPR/Cas9 Next-Generation Sequencing.

ABSTRACTCurrently available methods for the laboratory investigation of Legionella pneumophila outbreaks require organism culture. The ability to sequence L. pneumophila directly from clinical samples would significantly reduce delays. Here, we develop a method for targeted next-generation sequencing (NGS) of selected L. pneumophila genes utilizing a CRISPR/Cas9-based target enrichment system. We determine the method’s utility by typing cultured L. pneumophila isolates and subsequently apply the method directly to patient samples. We sequenced 10 L. pneumophila isolates by 2 methods, (i) whole-genome sequencing (WGS) and (ii) targeted (CRISPR/Cas9-based) finding low-abundance sequences by hybridization (FLASH)-NGS, sequencing 57 selected genes. The targeted NGS of 57 genes was more efficient than WGS, and phylogenetic analysis of the 57 genes yielded the same classification of the L. pneumophila isolates as that based on analysis of whole-genome data. Furthermore, targeted NGS of L. pneumophila performed directly on patient respiratory samples correctly classified the patients according to their corresponding cultured isolates. This provides proof of concept that targeted NGS can be used to sequence L. pneumophila directly from patient samples. Studies on a larger number of patient samples will further validate this method. Nonetheless, CRISPR/Cas9 targeted NGS methods have the potential to be widely applicable to microbial-outbreak investigations in the future, particularly in the context of difficult and slow-growing organisms.IMPORTANCE The bacterium Legionella pneumophila is responsible for outbreaks of serious and life-threatening pneumonia called Legionnaires’ disease. There is a need for new molecular methods that allow investigation of Legionella outbreaks directly from patient samples, without the need for prior microbiological culture, which causes delays. Our study aims to address this problem. We have utilized a CRISPR/Cas9-based targeted next-generation sequencing (NGS) method that can be applied directly on human specimens. Furthermore, we show that analysis of the sequences of a small number of targeted genes offers the same classification of L. pneumophila as that based on data derived from the whole genome. Given the rising interest globally in sequencing pathogens directly from human samples, CRISPR/Cas9 targeted NGS methods have the potential to be widely applicable to microbial-outbreak investigations in the future, particularly in the context of difficult and slow-growing organisms.

The complex epistemological challenge of data curation in dietary metabarcoding: Comment on "The precautionary principle and dietary DNA metabarcoding: Commonly used abundance thresholds change ecological interpretation" by Littleford-Colquhoun et al. (2022).

In their article, Littleford-Colquhoun et al. (2022) advise against using arbitrary relative read abundance (RRA) thresholds (i.e., minimum sequence copy thresholds) for removing low-abundance sequences since they can increase false negative rates in dietary DNA metabarcoding data sets. The main criticisms presented against these widespread methods are that they (i) are arbitrary, often existing as standard values or defined based on researcher-selected delineations, (ii) are subjective, varying between studies and contexts, and, most problematically, (iii) result in the exclusion of true positives, particularly rarely consumed taxa, to the detriment of ecological insight. We commend the authors for presenting a refreshing and timely perspective on this often neglected topic, which is certainly in need of greater discussion following over a decade of significant advances in dietary metabarcoding. In this complex epistemological problem of false positives versus false negatives, we feel that several of the points raised deserve additional discussion. We address these aspects below, including measured approaches to data filtration and consistent representation of RRAs, and we welcome any further discourse to solidify or refute the concepts therein.

ReadBouncer: precise and scalable adaptive sampling for nanopore sequencing.

MotivationNanopore sequencers allow targeted sequencing of interesting nucleotide sequences by rejecting other sequences from individual pores. This feature facilitates the enrichment of low-abundant sequences by depleting overrepresented ones in-silico. Existing tools for adaptive sampling either apply signal alignment, which cannot handle human-sized reference sequences, or apply read mapping in sequence space relying on fast graphical processing units (GPU) base callers for real-time read rejection. Using nanopore long-read mapping tools is also not optimal when mapping shorter reads as usually analyzed in adaptive sampling applications.ResultsHere, we present a new approach for nanopore adaptive sampling that combines fast CPU and GPU base calling with read classification based on Interleaved Bloom Filters. ReadBouncer improves the potential enrichment of low abundance sequences by its high read classification sensitivity and specificity, outperforming existing tools in the field. It robustly removes even reads belonging to large reference sequences while running on commodity hardware without GPUs, making adaptive sampling accessible for in-field researchers. Readbouncer also provides a user-friendly interface and installer files for end-users without a bioinformatics background.Availability and implementationThe C++ source code is available at https://gitlab.com/dacs-hpi/readbouncer.Supplementary information Supplementary data are available at Bioinformatics online.

Modification of polyacrylate sorbent coatings with carbodiimide crosslinker chemistry for sequence-selective DNA extraction using solid-phase microextraction.

Selective DNA extraction is immensely useful for the isolation and detection of low-abundance sequences. Oligonucleotide-modified substrates are often used to capture sequences of interest for downstream analysis. In this study, we explore the chemical modification of commercial-available polyacrylate solid-phase microextraction fibers for selective DNA analysis using carbodiimide crosslinker chemistry. Reproducible modification conditions are found and the fibers were subsequently applied for selective DNA analysis. Several experimental parameters such as stir-rate, desorption time, and buffer-type are optimized. The developed method was able to selectively extract the target DNA sequence (260 bp) in the presence of 100-fold excess interfering salmon testes DNA.

FLASH: a next-generation CRISPR diagnostic for multiplexed detection of antimicrobial resistance sequences.

The growing prevalence of deadly microbes with resistance to previously life-saving drug therapies is a dire threat to human health. Detection of low abundance pathogen sequences remains a challenge for metagenomic Next Generation Sequencing (NGS). We introduce FLASH (Finding Low Abundance Sequences by Hybridization), a next-generation CRISPR/Cas9 diagnostic method that takes advantage of the efficiency, specificity and flexibility of Cas9 to enrich for a programmed set of sequences. FLASH-NGS achieves up to 5 orders of magnitude of enrichment and sub-attomolar gene detection with minimal background. We provide an open-source software tool (FLASHit) for guide RNA design. Here we applied it to detection of antimicrobial resistance genes in respiratory fluid and dried blood spots, but FLASH-NGS is applicable to all areas that rely on multiplex PCR.

Evidence of photoacclimatization at mesophotic depths in the coral-Symbiodinium symbiosis at Flower Garden Banks National Marine Sanctuary and McGrail Bank

Similar to shallower conspecifics, mesophotic scleractinian corals found at ~ 30–150 m depths maintain important symbioses with photosynthetic microalgae in the genus Symbiodinium. Despite the importance of coral-algal symbioses in corals’ ability to thrive in multiple dynamic environments and potential role in connectivity, few studies have focused on mesophotic Symbiodinium assemblages. This study examines these assemblages in Montastraea cavernosa found at shallow (20–25 m) and mesophotic (45–50 m) depths at Flower Garden Banks National Marine Sanctuary and McGrail Bank, in the northwest Gulf of Mexico. Mesophotic corals contained significantly more Symbiodinium cells, more chlorophyll a per Symbiodinium cell, and more chlorophyll a and c2 per unit area coral tissue than shallow corals. However, both mesophotic and shallow M. cavernosa contained similar chlorophyll c2 per Symbiodinium cell. Next-generation sequencing of the internal transcribed spacer region (ITS2) of the ribosomal DNA indicated similar Symbiodinium assemblage diversity at all banks and between depths. All assemblages were dominated by sequences most closely related to S. goreaui, type C1, with three additional low-abundance sequences, identified as 2 C types and 1 A type, also consistently observed among colonies. Both the dominant C1 sequence and the background sequences persisted over two sampling years. These results suggest that algal symbiont assemblages will not limit connectivity potential in M. cavernosa in the northwest Gulf of Mexico. Furthermore, we hypothesize that increased Symbiodinium abundance may represent an effective light-harvesting strategy on light-limited mesophotic coral reefs.

AmpliconDuo: A Split-Sample Filtering Protocol for High-Throughput Amplicon Sequencing of Microbial Communities.

High throughput sequencing (HTSeq) of small ribosomal subunit amplicons has the potential for a comprehensive characterization of microbial community compositions, down to rare species. However, the error-prone nature of the multi-step experimental process requires that the resulting raw sequences are subjected to quality control procedures. These procedures often involve an abundance cutoff for rare sequences or clustering of sequences, both of which limit genetic resolution. Here we propose a simple experimental protocol that retains the high genetic resolution granted by HTSeq methods while effectively removing many low abundance sequences that are likely due to PCR and sequencing errors. According to this protocol, we split samples and submit both halves to independent PCR and sequencing runs. The resulting sequence data is graphically and quantitatively characterized by the discordance between the two experimental branches, allowing for a quick identification of problematic samples. Further, we discard sequences that are not found in both branches (“AmpliconDuo filter”). We show that the majority of sequences removed in this way, mostly low abundance but also some higher abundance sequences, show features expected from random modifications of true sequences as introduced by PCR and sequencing errors. On the other hand, the filter retains many low abundance sequences observed in both branches and thus provides a more reliable census of the rare biosphere. We find that the AmpliconDuo filter increases biological resolution as it increases apparent community similarity between biologically similar communities, while it does not affect apparent community similarities between biologically dissimilar communities. The filter does not distort overall apparent community compositions. Finally, we quantitatively explain the effect of the AmpliconDuo filter by a simple mathematical model.

Optimizing a PCR protocol for cpn60-based microbiome profiling of samples variously contaminated with host genomic DNA.

BackgroundThe current recommended protocol for chaperonin-60 (cpn60) universal target based microbiome profiling includes universal PCR of microbiome samples across an annealing temperature gradient to maximize the diversity of sequences amplified. However, the value of including this gradient approach has not been formally evaluated since the optimization of a modified universal PCR primer cocktail for cpn60 PCR. PCR conditions that maximize representation of the microbiome while minimizing PCR-associated distortion of the community structure, especially in samples containing large amounts of host genomic DNA are critical. The goal of this study was to measure the effects of PCR annealing temperature and the ratio of host to bacterial DNA on the outcome of microbiota analysis, using pig microbiota as a model environment.FindingsSix samples were chosen with an anticipated range of ratios of pig to bacterial genomic DNA, and universal cpn60 PCR amplification with an annealing temperature gradient was used to create libraries for pyrosequencing, resulting in 426,477 sequences from the six samples. The sequences obtained were classified as target (cpn60) or non-target based on the percent identity of their closest match to the cpnDB reference database, and target sequences were further processed to create microbiome profiles for each sample at each annealing temperature. Annealing temperature affected the amount of PCR product generated, with more product generated at higher temperatures. Samples containing proportionally more host genomic DNA yielded more non-target reads, especially at lower annealing temperatures. However, microbiome composition for each sample across the annealing temperature gradient remained consistent at both the phylum and operational taxonomic unit levels. Although some microbial sequences were detected at only one annealing temperature, these sequences accounted for a minority of the total microbiome.ConclusionsThese results indicate that PCR annealing temperature does have an affect on cpn60 based microbiome profiles, but that most of the differences are due to differences in detection of low abundance sequences. Higher annealing temperatures resulted in larger amounts of PCR product and lower amounts of non-target sequence amplification, especially in samples containing proportionally large amounts of host DNA. Taken together these results provide important information to guide decisions about experimental design for cpn60 based microbiome studies.Electronic supplementary materialThe online version of this article (doi:10.1186/s13104-015-1170-4) contains supplementary material, which is available to authorized users.

Influence of artifact removal on rare species recovery in natural complex communities using high-throughput sequencing.

Large-scale high-throughput sequencing techniques are rapidly becoming popular methods to profile complex communities and have generated deep insights into community biodiversity. However, several technical problems, especially sequencing artifacts such as nucleotide calling errors, could artificially inflate biodiversity estimates. Sequence filtering for artifact removal is a conventional method for deleting error-prone sequences from high-throughput sequencing data. As rare species represented by low-abundance sequences in datasets may be sensitive to artifact removal process, the influence of artifact removal on rare species recovery has not been well evaluated in natural complex communities. Here we employed both internal (reliable operational taxonomic units selected from communities themselves) and external (indicator species spiked into communities) references to evaluate the influence of artifact removal on rare species recovery using 454 pyrosequencing of complex plankton communities collected from both freshwater and marine habitats. Multiple analyses revealed three clear patterns: 1) rare species were eliminated during sequence filtering process at all tested filtering stringencies, 2) more rare taxa were eliminated as filtering stringencies increased, and 3) elimination of rare species intensified as biomass of a species in a community was reduced. Our results suggest that cautions be applied when processing high-throughput sequencing data, especially for rare taxa detection for conservation of species at risk and for rapid response programs targeting non-indigenous species. Establishment of both internal and external references proposed here provides a practical strategy to evaluate artifact removal process.

12-OR

Aim Genotyping of DRB1 and DRB3/4/5 is performed in our lab using generic DRB PCR primers, followed by sequencing on the 454 GS FLX and use of Conexio Genomics software. During genotyping homozygous cell lines, we detected additional sequences at about 7% of the read depth of the reported true allele. These artifacts appeared to be PCR crossovers between the DRB1 and DRB3/4/5 sequences. The crossovers usually had one or more mismatches with the IMGT database but occasionally matched a very rare allele (e.g. DRB1 ∗ 03:42). We performed a study to determine if these sequences were in vitro artifacts and to determine if some rare DRB alleles in the IMGT database might, in fact, be in vitro artifacts that had been submitted. Methods To test for in vitro artifact formation during late cycles of genomic PCR, we amplified the DRB loci from 21 homozygous cells lines under both our standard PCR conditions (35 cycles) and a reduced number of cycles (28), followed by 454 sequencing. Using the same PCR conditions, we also sequenced the DRB loci of four samples, each of which had previously been identified as having a novel DRB allele that was subsequently submitted to the IMGT database. Results For all homozygous 21 cell lines, amplification for 35 cycles gave low abundance sequences as possible “second alleles.” In 36% of the cases, these artifactual sequences corresponded to named alleles. Such sequences were not detected with amplification for 28 cycles. In the case of 4 samples with rare alleles previously submitted to the IMGT, we found that 3 were true alleles. The fourth, while not a crossover product, contained a sequencing error which was revealed by our 454 HLA genotyping assay. Conclusions Formation of in vitro crossover products of the DRB loci can occur in late cycles of PCR. PCR conditions have been identified that minimize generation of these in vitro artifacts. Clonal sequencing on the 454 GS FLX is a valuable method for revealing both these artifacts and traditional sequencing errors.

Genovo: De Novo Assembly for Metagenomes

Next-generation sequencing technologies produce a large number of noisy reads from the DNA in a sample. Metagenomics and population sequencing aim to recover the genomic sequences of the species in the sample, which could be of high diversity. Methods geared towards single sequence reconstruction are not sensitive enough when applied in this setting. We introduce a generative probabilistic model of read generation from environmental samples and present Genovo, a novel de novo sequence assembler that discovers likely sequence reconstructions under the model. A nonparametric prior accounts for the unknown number of genomes in the sample. Inference is performed by applying a series of hill-climbing steps iteratively until convergence. We compare the performance of Genovo to three other short read assembly programs in a series of synthetic experiments and across nine metagenomic datasets created using the 454 platform, the largest of which has 311k reads. Genovo's reconstructions cover more bases and recover more genes than the other methods, even for low-abundance sequences, and yield a higher assembly score. Supplementary Material is available at www.liebertoinline.com/cmb .

Human Cdc5, a regulator of mitotic entry, can act as a site-specific DNA binding protein.

G(2)/M progression requires coordinated expression of many gene products, but little is known about the transcriptional regulators involved. We recently identified human Cdc5, a positive regulator of G(2)/M in mammalian cells. We also demonstrated the presence of a latent activation domain in its carboxyl terminus, suggesting that human Cdc5 regulates G(2)/M through transcriptional activation. Despite the presence of a DNA binding domain, studies by others have failed to identify a preferential binding site for Cdc5 family members. In addition, Cdc5 recently has been associated with the splicesome in several organisms, suggesting that it may not act through DNA binding. We now report the identification of a 12 bp sequence to which human Cdc5 binds specifically and with high affinity through its amino terminus. We show that this DNA-protein interaction is capable of activating transcription. We also used a selection system in yeast to identify human genomic fragments that interact with human Cdc5. Several of these contained sequences similar to the binding site. We demonstrate that these bind human Cdc5 with similar specificity and affinity. These experiments provide the first evidence that Cdc5 family members can act as site-specific DNA binding proteins, and that human Cdc5 may interact with specific, low abundance sequences in the human genome. This raises the possibility that Cdc5 proteins may participate in more than one process necessary for regulated cell division.

PCR-generated cDNA libraries from reduced numbers of mouse oocytes.

We describe a rapid and reproducible method for cloning cDNA amplified from 10 mouse oocytes. The procedure consists in priming cDNA synthesis from a crude cellular extract using an oligo d(T) containing primer and submitting the size-limited cDNA first strand to poly(dG) tailing. The whole cDNA population is then polymerase chain reaction (PCR) amplified using two primers complementary to oligo d(A) and oligo d(G) ends of the cDNA. In this procedure no purification steps are required. We obtained about 5 x 10(6) clones from 10 oocytes. Screening of the library showed that the relative abundance of the transcripts was preserved during amplification and cloning and that the procedure allows cloning of low-abundance sequences at least as rare as 0.008% of the mRNA. The repeatable generation of representative cDNA libraries from reduced numbers of oocytes or embryos should open new opportunities for obtaining genetic information from mammalian preimplantation embryos.

Rapid identification of hybridization probes for chromosomal walking

The presence of repeated elements in restriction fragments used as hybridization probes for chromosomal walking poses a major obstacle to the success of this gene-cloning strategy. This report describes a simple and rapid means of identifying restriction fragments devoid of repeated sequences and therefore useful as hybridization probes for chromosomal walking. Restriction fragments derived from a genomic DNA clone are Southern blotted and hybridized to nick-translated total genomic [ 32P]DNA. Fragments of the genomic clone that contain high abundance sequences (i.e., repeated elements) hybridize strongly to their nick-translated counterparts, which, due to their high copy number, comprise a significant proportion of the total genomic DNA probe. Conversely, fragments containing single-copy or low-abundance sequences do not hybridize, as their nick-translated counterparts are poorly represented in the total genomic DNA probe. These latter fragments, by virtue of their low-abundance sequences, are well suited as probes for chromosomal walking. Ensuring the absence of repeated elements in restriction fragments prior to their purification and utilization as chromosomal walking probes results in marked savings of time, effort and materials.

The effect of temperature change on gene expression in Paramecium primaurelia

When paramecia grown at 24°C are transferred rapidly to 32°C, DNA and protein synthesis continue uninterrupted but at higher rates. Electron microscopic observations indicate that more of the macronuclear chromatin is transcribed at the elevated temperature. This interpretation is supported by hybridization experiments which show that the percentage of the macronuclear genome transcribed into poly(A) + RNA is 24°C and 35% at 32°C. Kinetic analysis of cDNA-poly(A) + RNA hybridizations reveals three abundance classes of poly(A) + RNA and indicates that the number of genes expressing low abundance sequences is about 9000 at 24°C and 13000 at 32°C. The intermediately abundant and highly abundant classes are represented by 100–200 and 1–3 different kinds of RNA sequence, respectively. Cross hybridization shows that changes occur throughout the distribution of abundance classes of poly(A) + RNA with increase in temperature.

Molecular cloning of developmentally regulated, low-abundance mRNA sequences from embryonic muscle.

To study the role of low-abundance, embryonic muscle-specific gene transcripts, we have developed a method to screen cDNA clones from embryonic muscle for such sequences. The protocol involves two stages: first, partial enrichment for cDNA clones carrying possible embryo-specific sequences by selecting clones of low-abundance sequences; and second, determination, by hybridization to RNA attached to diazobenzyloxymethyl-paper, which sequences from this category are regulated in an embryonic muscle-specific manner during development. At least three different clones were obtained which hybridized to sequences present in early muscle development but absent, or present at relatively low levels, at late embryonic and adult muscle stages. Two of these clones were not muscle-specific because they hybridized to poly(A)+RNA from liver or brain or both. The third clone, 106A4, did not detectably hybridize to total poly(A)+RNA at any stage of brain or liver development tested. This sequence also was not detectable in poly(A)+RNA from embryonic muscle progenitor cells. Thus, the 106A4 sequence is a likely candidate for an embryonic muscle-specific sequence. We have demonstrated that the 106A4 sequence is a mRNA, although the specific identity and function of the translated product is unknown. The method used to identify embryonic muscle-specific cDNA clones should be generally applicable for obtaining clones for low abundance transcripts regulated in a tissue-specific or developmental stage-specific manner.