The BabyBio Dsalt 1 ml and 5 ml desalting columns remove salt essentially completely from BSA samples up to 100 µl and 750 µl respectively, with partial removal of 82% and 93% at larger loads. Connecting columns in series scales the working volume, and three 5 ml columns in series buffer-exchanged…
BabyBio Dsalt 1 ml and 5 ml columns are prepacked desalting columns for protein solutions. On the 1 ml format, a 2 mg/ml bovine serum albumin BSA solution in 25 mM sodium phosphate, 500 mM NaCl, pH 7.0 is essentially completely desalted at sample volumes of 20 µl and 100 µl; at 300 µl, salt removal falls to 82%. On the 5 ml format, the same sample is essentially completely desalted at 100 µl and 750 µl, and a 1.5 ml sample still loses 93% of its salt. The columns can be connected in series: a 1 ml sample run over five 1 ml columns matched one 5 ml column, and three 5 ml columns in series buffer-exchanged a 3 ml pool of human polyclonal IgG eluted at pH 2.7 from a Protein A column, restoring neutral pH and preventing acid-induced aggregation. All of the column performance figures come from the manufacturer's application note, and none has been independently reproduced in the peer-reviewed literature.
Desalting columns work by size-exclusion chromatography, not by binding. The prepacked medium is a gel with pores large enough to admit salt ions and small molecules but small enough to exclude proteins above a cutoff. Excluded proteins travel only through the liquid between the beads, reach the column outlet first, and elute as a compact band. Salt and other low-molecular-weight solutes diffuse into the pores, follow a longer path, and emerge later as a separate peak. One pass therefore replaces the sample's high-salt buffer with the column's equilibration buffer while the original salt leaves in the trailing fraction.
Two detectors make the separation legible. UV absorbance at 280 nm marks the protein band; conductivity marks the salt band. Because salt elutes after protein, the collected protein fraction must be cut before the conductivity signal rises. A fraction collector triggered by conductivity alone would collect the salt peak, so the cut is best made from the A280 trace or a calibrated delay.
The sample-volume limits follow from the same mechanism. Protein and salt each occupy a band of liquid moving down the bed. At small sample volumes the bands are fully separated. As the applied volume grows, the trailing edge of the protein band catches the leading edge of the salt band, and salt begins to co-elute with the protein. Desalting is a coarse group separation; its working range is set by sample volume relative to bed volume, not by selective binding.
One nuance matters when reading the vendor data. The equilibration and elution buffer in the desalting runs was 25 mM sodium phosphate, 150 mM NaCl, pH 7.0, while the BSA sample was in 25 mM sodium phosphate, 500 mM NaCl, pH 7.0. Complete salt removal therefore means the protein finishes in the 150 mM NaCl running buffer, not in salt-free water. Desalting columns are buffer-exchange devices; the salt load removed is the excess NaCl over the target buffer.
The performance figures below come from a single application note authored by Kajsa Eriksson, Johanna Tawe, and Lars Haneskog of Bio-Works in Uppsala, Sweden, and published Jun 10, 2025. The test conditions were identical for both formats: BSA at 2 mg/ml in 25 mM sodium phosphate, 500 mM NaCl, pH 7.0, with detection by A280 for protein and conductivity for salt.
| Column | Sample volume | Salt removed | Peak separation |
|---|---|---|---|
| BabyBio Dsalt 1 ml | 20 µl | Essentially complete | Resolved |
| BabyBio Dsalt 1 ml | 100 µl | Essentially complete | Resolved |
| BabyBio Dsalt 1 ml | 300 µl | 82% | Partial overlap |
| BabyBio Dsalt 5 ml | 100 µl | Essentially complete | Resolved |
| BabyBio Dsalt 5 ml | 750 µl | Essentially complete | Resolved |
| BabyBio Dsalt 5 ml | 1.5 ml | 93% | Partial overlap |
Within the fully resolved range, the protein and salt peaks were separated from each other, so the protein fraction carried no detectable salt by conductivity. At the larger volumes the peaks overlapped, and the manufacturer quantified the salt remaining with the protein: 82% removed at 300 µl on the 1 ml column, 93% removed at 1.5 ml on the 5 ml column.
The two formats behave consistently when sample volume is expressed relative to bed volume. The fully resolved volumes, 100 µl on 1 ml and 750 µl on 5 ml, are 10% and 15% of the bed respectively. The overloaded volumes, 300 µl on 1 ml and 1.5 ml on 5 ml, are both 30% of the bed. At that same relative load the 5 ml column retained more salt 93% removed versus 82% , which suggests the two formats differ in bed geometry, such as bed length, details the manufacturer does not disclose.
The limits of this dataset are clear. Only BSA was tested, at a single concentration, in a single buffer pair. Recovery of protein, the behavior of viscous, particulate, or high-concentration samples, and the performance of the columns with antibodies or other proteins under these conditions were not reported. BSA is a convenient and forgiving test protein; other proteins may aggregate, adsorb to the medium, or interact with it differently.
The same size-exclusion logic that sets the volume limits also enables scale-up. Each column added in series contributes more bed volume and another full separation path, so a bank of small columns can process a sample that would exceed a single column's range. The manufacturer tested this directly: a 1 ml BSA sample run on five BabyBio Dsalt 1 ml columns connected in series gave results comparable to the same sample on a single 5 ml column. The two arrangements are matched in total bed volume, 5 ml in both cases, so the comparable outcome is what the mechanism predicts.
For the user, the rule is to keep sample volume proportional to total connected bed volume, with the same absolute limit used for a single column of that combined size. A 1 ml sample on five 1 ml columns behaves like a 1 ml sample on one 5 ml column. The approach extends to bigger trains by the same arithmetic. The trade-offs are longer run times, higher back pressure, and more elution volume, none of which the application note quantifies. The manufacturer states that the columns connect directly to most chromatography systems, and the series runs in its note are consistent with that claim, but no independent test of the connectors appears in the literature.
What the series experiment does not show is a maximum train length. It shows one configuration, five 1 ml columns, at one sample volume. Whether ten or twenty columns in series preserve the same resolution is untested. The monotonic improvement in separation with path length suggests the approach should hold until pressure or dilution becomes limiting, but that is inference from the mechanism, not a measured result.
Protein A affinity purification works by binding antibodies at neutral pH and releasing them at low pH, typically around pH 2.7 to 3. The released antibody therefore enters a solution far below its stable pH range. At that acidity, antibodies partially unfold: buried hydrophobic surfaces become exposed, and the partially unfolded molecules collide and aggregate. The standard countermeasure is to neutralize the eluate as quickly as possible. Desalting offers a one-step version of this: a column equilibrated in neutral buffer both strips the acidic eluent and returns the antibody to a near-neutral environment.
| Workflow step | Buffer and conditions | Volume |
|---|---|---|
| Load | Human polyclonal IgG at 1 mg/ml in PBS, pH 7.4 | 20 ml |
| Elute | 100 mM Gly-HCl, pH 2.7 | 3 ml |
| Buffer exchange | Three BabyBio Dsalt 5 ml columns in series | 3 ml pool |
The demonstration used a 20 ml load of human polyclonal IgG at 1 mg/ml in PBS, pH 7.4 on a BabyBio A 1 ml affinity column. Elution with 100 mM Gly-HCl, pH 2.7 produced a 3 ml pool, which was applied immediately to three BabyBio Dsalt 5 ml columns in series. The train restored neutral pH, and the manufacturer reports that the antibody was protected from acid-induced aggregation. The configuration keeps the sample volume at 20% of the total 15 ml bed volume, below the partial-overlap condition seen at 30% in the BSA data.
The workflow is plausible and the configuration is sound, but the evidence is one demonstration with one antibody type at one concentration. No aggregation measurements, such as monomer recovery by size-exclusion HPLC or particle counts by dynamic light scattering, appear in the public description. The protective effect is attributed to the fast return to neutral pH, which is consistent with the aggregation mechanism, but "prevented aggregation" is the manufacturer's conclusion, not a measured value. Different isotypes, higher concentrations, and harsher elution conditions all raise the aggregation risk and would need their own verification.
No indexed study has independently evaluated BabyBio Dsalt columns. The salt-removal figures, the series comparison, and the antibody neutralization run are the manufacturer's own measurements. That leaves the burden of verification on the user, and it makes the columns' stated performance a vendor claim rather than an established fact.
The broader literature does show why desalting and buffer exchange are consequential steps in peptide and protein work. Reviews of kokumi-active peptide preparation identify extraction and enzymatic generation as the main routes to these peptides and describe evaluation through sensory panels and calcium-sensing receptor assays PMID 33103468 ; the buffer and salt state of the test sample is part of any such assay. Structural studies of peptide-bond isosteres, including thioamide, ester, alkene, and fluoroalkene replacements, hinge on comparing electronic, conformational, and noncovalent interaction properties PMID 21751326 , comparisons that require peptides in defined solution conditions. A statistical analysis of combinatorial peptide synthesis established that libraries do not contain equimolar amounts of each sequence and that bead numbers and design determine the spread of relative concentrations PMID 8789352 , a result that makes post-synthesis handling, including salt and buffer control, essential before screening data can be interpreted.
Applied studies make the same point. A bivalent EGFR-binding peptide conjugated to a photosensitizer reduced viability of EGFR-positive cells in a light-dose- and drug-concentration-dependent manner PMID 38614014 , which depends on a correctly formulated conjugate. A review of antibacterial peptides notes that peptide drugs are less immunogenic than other biologics but can still cause allergic reactions, and recommends preclinical evaluation with cytotoxicity, basophil activation, and lymphocyte activation assays PMID 29443886 . Family-wide screens of RAPID ALKALINIZATION FACTOR peptides in Arabidopsis linked alkalinization and seedling growth inhibition to the receptor kinase FERONIA PMID 34608971 , results obtained with synthetic peptides delivered in defined buffers. In all of these settings, the salt, pH, and purity of the peptide solution modulate receptor engagement, aggregation, and assay readout, which is exactly why desalting and buffer exchange sit between preparation and evaluation. None of these studies tests the Bio-Works columns, so they contextualize the workflow without validating the product.
Practical guidance follows directly from the data.
The open questions are substantial. The maximum sample volume that still gives complete salt removal is unknown for both formats; the data show complete removal at 100 µl and 750 µl and partial removal at 300 µl and 1.5 ml, with no measurements between. Protein recovery after desalting is unreported. Column lifetime, cleaning, and any decline with repeated use are unreported. Everything was measured with BSA, so performance with antibodies, glycoproteins, high-viscosity samples, or complex mixtures is untested. The neutralization demonstration used human polyclonal IgG only.
None of this means the columns do not work as described. The manufacturer's figures are internally consistent and mechanistically sensible: complete salt removal within the stated volume ranges, partial removal above them, series scale-up that matches a single larger column, and a neutralization run sized to stay within the working envelope. What the figures do not support is any claim beyond those tested conditions.
The most useful next step for a potential buyer is to run their own protein in their own buffer through one column, measuring A280 and conductivity, before committing to a desalting step at scale. Desalting performance is easy to verify, and the verification takes less time than the purification runs that depend on it.
Related reading: Combination peptide therapy may reduce opioid side effects, Prepacked OptioBio Columns for Small-Scale IEX and Process Development, CAPA tracks biotherapeutic cell uptake and organelle distribution, Cyclic Hexapeptide Mimics Disrupt HIV-1 Integrase and LEDGF/p75.